WebKit-https.git
5 years ago[JSC] Speed up URL encode/decode by using bitmaps instead of strchr().
akling@apple.com [Mon, 18 May 2015 17:46:43 +0000 (17:46 +0000)]
[JSC] Speed up URL encode/decode by using bitmaps instead of strchr().
<https://webkit.org/b/145115>

Reviewed by Anders Carlsson.

We were calling strchr() for every character when doing URL encoding/decoding and it stood out
like a sore O(n) thumb in Instruments. Optimize this by using a Bitmap<256> instead.

5.5% progression on Kraken/stanford-crypto-sha256-iterative.

* runtime/JSGlobalObjectFunctions.cpp:
(JSC::makeCharacterBitmap):
(JSC::encode):
(JSC::decode):
(JSC::globalFuncDecodeURI):
(JSC::globalFuncDecodeURIComponent):
(JSC::globalFuncEncodeURI):
(JSC::globalFuncEncodeURIComponent):
(JSC::globalFuncEscape):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184501 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoNetwork Cache: Make Data::adoptMap take the ownership of the file descriptor
carlosgc@webkit.org [Mon, 18 May 2015 06:31:25 +0000 (06:31 +0000)]
Network Cache: Make Data::adoptMap take the ownership of the file descriptor
https://bugs.webkit.org/show_bug.cgi?id=144950

Reviewed by Antti Koivisto.

It will be required to implement ShareableResource for soup
network backend. Also move the common code of Data to a common
file and add mapToFile(). The mapFile version that receives a file
descriptor has been renamed to adoptAndMapFile().

* CMakeLists.txt: Add new file to compilation.
* NetworkProcess/cache/NetworkCacheBlobStorage.cpp:
(WebKit::NetworkCache::BlobStorage::add): Use mapToFile().
* NetworkProcess/cache/NetworkCacheData.cpp: Added.
(WebKit::NetworkCache::Data::mapToFile): Write the data to the
given file and map it.
(WebKit::NetworkCache::mapFile):
(WebKit::NetworkCache::adoptAndMapFile):
(WebKit::NetworkCache::computeSHA1):
(WebKit::NetworkCache::bytesEqual):
* NetworkProcess/cache/NetworkCacheData.h:
* NetworkProcess/cache/NetworkCacheDataCocoa.mm:
(WebKit::NetworkCache::Data::adoptMap): Close the file descriptor.
* NetworkProcess/cache/NetworkCacheDataSoup.cpp:
(WebKit::NetworkCache::Data::Data): Use a constructor that
receives a file descriptor instead of the one receiving
Backing. If the file descriptor is not -1 then the Data is a map.
(WebKit::NetworkCache::MapWrapper::~MapWrapper): Also close the
file descriptor.
(WebKit::NetworkCache::Data::adoptMap): Pass the file descriptor
to the MapWrapper and create the Data passing the file descriptor.
* UIProcess/API/APIUserContentExtensionStore.cpp:
(API::openAndMapContentExtension): Use mapFile that receives a
file path.
(API::compiledToFile): Use adoptAndMapFile() and don't close the
descriptor.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184449 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoDo not use fastMallocGoodSize anywhere
benjamin@webkit.org [Mon, 18 May 2015 06:26:10 +0000 (06:26 +0000)]
Do not use fastMallocGoodSize anywhere
https://bugs.webkit.org/show_bug.cgi?id=145103

Reviewed by Michael Saboff.

Source/JavaScriptCore:

* assembler/AssemblerBuffer.h:
(JSC::AssemblerData::AssemblerData):
(JSC::AssemblerData::grow):

Source/WTF:

It is silly we see fastMallocGoodSize in profiles, it does absolutely nothing.

This patch keeps fastMallocGoodSize() around for older code linking
with newer WebKit, but remove any use of it inside WebKit.

* wtf/FastMalloc.cpp:
(WTF::fastMallocGoodSize):
* wtf/FastMalloc.h:
* wtf/Vector.h:
(WTF::VectorBufferBase::allocateBuffer):
(WTF::VectorBufferBase::tryAllocateBuffer):
(WTF::VectorBufferBase::reallocateBuffer):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184448 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[JSC] Make StringRecursionChecker faster in the simple cases without any recursion
benjamin@webkit.org [Mon, 18 May 2015 06:23:31 +0000 (06:23 +0000)]
[JSC] Make StringRecursionChecker faster in the simple cases without any recursion
https://bugs.webkit.org/show_bug.cgi?id=145102

Reviewed by Darin Adler.

Source/JavaScriptCore:

In general, the array targeted by Array.toString() or Array.join() are pretty
simple. In those simple cases, we spend as much time in StringRecursionChecker
as we do on the actual operation.

The reason for this is the HashSet stringRecursionCheckVisitedObjects used
to detect recursion. We are constantly adding and removing objects which
dirty buckets and force constant rehash.

This patch adds a simple shortcut for those simple case: in addition to the HashSet,
we keep a pointer to the root object of the recursion.
In the vast majority of cases, we no longer touch the HashSet at all.

This patch is a 12% progression on the overall score of ArrayWeighted.

* runtime/StringRecursionChecker.h:
(JSC::StringRecursionChecker::performCheck):
(JSC::StringRecursionChecker::~StringRecursionChecker):
* runtime/VM.h:

LayoutTests:

Improve the coverage a tiny bit.

* js/array-string-recursion-expected.txt: Added.
* js/array-string-recursion.html: Added.
* js/script-tests/array-string-recursion.js: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184447 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[CSS Grid Layout] Add scrollbar width in intrinsic logical widths computation
rego@igalia.com [Mon, 18 May 2015 06:16:58 +0000 (06:16 +0000)]
[CSS Grid Layout] Add scrollbar width in intrinsic logical widths computation
https://bugs.webkit.org/show_bug.cgi?id=145021

Source/WebCore:

Like for flexboxes we've to take into account the scrollbar logical
width while computing the intrinsic min and max logical widths.

Reviewed by Sergio Villar Senin.

Test: fast/css-grid-layout/compute-intrinsic-widths-scrollbar.html

* rendering/RenderGrid.cpp:
(WebCore::RenderGrid::computeIntrinsicLogicalWidths): Add scrollbar
logical width.

LayoutTests:

Reviewed by Sergio Villar Senin.

* fast/css-grid-layout/compute-intrinsic-widths-scrollbar-expected.txt: Added.
* fast/css-grid-layout/compute-intrinsic-widths-scrollbar.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184446 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoInsert store barriers late so that IR transformations don't have to worry about them
fpizlo@apple.com [Mon, 18 May 2015 03:39:28 +0000 (03:39 +0000)]
Insert store barriers late so that IR transformations don't have to worry about them
https://bugs.webkit.org/show_bug.cgi?id=145015

Reviewed by Geoffrey Garen.

We have had three kinds of bugs with store barriers. For the sake of discussion we say
that a store barrier is needed when we have something like:

    base.field = value

- We sometimes fail to realize that we could remove a barrier when value is a non-cell.
  This might happen if we prove value to be a non-cell even though in the FixupPhase it
  wasn't predicted non-cell.

- We sometimes have a barrier in the wrong place after object allocation sinking. We
  might sink an allocation to just above the store, but that puts it just after the
  StoreBarrier that FixupPhase inserted.

- We don't remove redundant barriers across basic blocks.

This comprehensively fixes these issues by doing store barrier insertion late, and
removing the store barrier elision phase. Store barrier insertion uses an epoch-based
algorithm to determine when stores need barriers. Briefly, a barrier is not needed if
base is in the current GC epoch (i.e. was the last object that we allocated or had a
barrier since last GC) or if base has a newer GC epoch than value (i.e. value would have
always been allocated before base). We do conservative things when merging epoch state
between basic blocks, and we only do such inter-block removal in the FTL. FTL also
queries AI to determine what type we've proved about value, and avoids barriers when
value is not a cell. FixupPhase still inserts type checks on some stores, to maximize
the likelihood that this AI-based removal is effective.

Rolling back in after fixing some debug build test failures.

* CMakeLists.txt:
* JavaScriptCore.vcxproj/JavaScriptCore.vcxproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* dfg/DFGBlockMap.h:
(JSC::DFG::BlockMap::at):
* dfg/DFGConstantFoldingPhase.cpp:
(JSC::DFG::ConstantFoldingPhase::emitPutByOffset):
* dfg/DFGEpoch.h:
(JSC::DFG::Epoch::operator<):
(JSC::DFG::Epoch::operator>):
(JSC::DFG::Epoch::operator<=):
(JSC::DFG::Epoch::operator>=):
* dfg/DFGFixupPhase.cpp:
(JSC::DFG::FixupPhase::fixupNode):
(JSC::DFG::FixupPhase::speculateForBarrier):
(JSC::DFG::FixupPhase::insertStoreBarrier): Deleted.
* dfg/DFGPlan.cpp:
(JSC::DFG::Plan::compileInThreadImpl):
* dfg/DFGStoreBarrierElisionPhase.cpp: Removed.
* dfg/DFGStoreBarrierElisionPhase.h: Removed.
* dfg/DFGStoreBarrierInsertionPhase.cpp: Added.
(JSC::DFG::performFastStoreBarrierInsertion):
(JSC::DFG::performGlobalStoreBarrierInsertion):
* dfg/DFGStoreBarrierInsertionPhase.h: Added.
* ftl/FTLOperations.cpp:
(JSC::FTL::operationMaterializeObjectInOSR): Fix an unrelated debug-only bug.
* tests/stress/load-varargs-then-inlined-call-and-exit.js: Test for that debug-only bug.
* tests/stress/load-varargs-then-inlined-call-and-exit-strict.js: Strict version of that test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184445 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[Streams API] Delegate ReadableStreamReader reference counting to ReadableStream
youenn.fablet@crf.canon.fr [Sun, 17 May 2015 18:44:39 +0000 (18:44 +0000)]
[Streams API] Delegate ReadableStreamReader reference counting to ReadableStream
https://bugs.webkit.org/show_bug.cgi?id=144907

Reviewed by Darin Adler.

Changed the link between readadable stream and reader.
Controller ref()/deref() now increments/decrements its stream ref counter, similarly to ReadableStreamController.
This ensures that even if JS scripts do not keep track of the readable stream,
the readable stream will not be disposed as long as the JS script has access to its reader.

All readers of a given stream are kept by the stream, either in an array (for released readers)
or and in ReadableStream::m_reader for the active reader.
This removes the need for the code synchronizing stream and reader.

As a reader can now already count on its stream, the reader no longer needs to store the error in errored state.
Removal of ReadableJSStream::Reader as closed promise rejection error is directly retrieved from the ReadableStream.
Moved the creation of reader directly in ReadableStream.

Next step should be to remove ReadableStreamReader::m_state and to delegate the handling of ReadableStreamReader closed promise callbacks to ReadableStream.

No change in behavior.

* Modules/streams/ReadableStream.cpp:
(WebCore::ReadableStream::changeStateToClosed):
(WebCore::ReadableStream::changeStateToErrored):
(WebCore::ReadableStream::getReader):
* Modules/streams/ReadableStream.h:
(WebCore::ReadableStream::reader):
* Modules/streams/ReadableStreamReader.cpp:
(WebCore::ReadableStreamReader::clean):
(WebCore::ReadableStreamReader::ref):
(WebCore::ReadableStreamReader::deref):
(WebCore::ReadableStreamReader::closed):
(WebCore::ReadableStreamReader::changeStateToClosed):
(WebCore::ReadableStreamReader::changeStateToErrored):
* Modules/streams/ReadableStreamReader.h:
(WebCore::ReadableStreamReader::ReadableStreamReader):
* Modules/streams/ReadableStreamReader.idl:
* bindings/js/JSReadableStreamCustom.cpp:
(WebCore::JSReadableStream::getReader):
* bindings/js/JSReadableStreamReaderCustom.cpp:
(WebCore::JSReadableStreamReader::closed):
(WebCore::constructJSReadableStreamReader):
* bindings/js/ReadableJSStream.cpp:
(WebCore::ReadableJSStream::storeError):
(WebCore::ReadableJSStream::ReadableJSStream): Deleted.
* bindings/js/ReadableJSStream.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184444 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoCrash when uploading huge files to YouTube or Google Drive
ap@apple.com [Sun, 17 May 2015 17:39:15 +0000 (17:39 +0000)]
Crash when uploading huge files to YouTube or Google Drive
https://bugs.webkit.org/show_bug.cgi?id=145083
rdar://problem/15468529

Reviewed by Darin Adler.

This fixes the crash, but uploading will fail.

* fileapi/FileReaderLoader.cpp:
(WebCore::FileReaderLoader::start): Tell SubresourceLoader to not store a copy of
all received data, FileReaderLoader has its own buffer.
(WebCore::FileReaderLoader::didReceiveResponse): Fixed a bounds check - not every
64-bit value that doesn't fit into 32 bits is negative. With this, FileReader fails
on huge files right away, as intended.
(WebCore::FileReaderLoader::didReceiveData): Fixed multiple bugs in code that's
executed when size is not available upfront. This is the code that used to crash,
but with the above fix, it's not executed by YouTube.
Not only overflow was handled incorrectly, but even simply growing a buffer for
append was buggy.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184443 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agobisect-builds: Add 'retry' option when prompting whether the bug reproduced
ddkilzer@apple.com [Sun, 17 May 2015 17:36:36 +0000 (17:36 +0000)]
bisect-builds: Add 'retry' option when prompting whether the bug reproduced
<http://webkit.org/b/145100>

Reviewed by Darin Adler.

* Scripts/bisect-builds:
- Add PROMPT_ANSWER_* constants for yes/no/retry/broken prompt.
- Switch to using PROMPT_RESPONSE_* constants when testing the
  value of $didReproduceBug.
- Add do { } while loops to implement 'retry' mode.
(promptForTest):
- Switch to returning PROMPT_RESPONSE_* constants.
- Add support for 'retry' response.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184442 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[iOS] WKPDFView should not follow javascript: links
aestes@apple.com [Sun, 17 May 2015 00:24:43 +0000 (00:24 +0000)]
[iOS] WKPDFView should not follow javascript: links
https://bugs.webkit.org/show_bug.cgi?id=145101

Reviewed by Dan Bernstein.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::navigateToPDFLinkWithSimulatedClick): Returned early if url's protocol is javascript:.
(WebKit::WebPageProxy::navigateToURLWithSimulatedClick): Renamed to navigateToPDFLinkWithSimulatedClick to
reflect that this function is exclusively for following PDF links.
* UIProcess/WebPageProxy.h:
* UIProcess/ios/WKPDFView.mm:
(-[WKPDFView annotation:wasTouchedAtPoint:controller:]):
(-[WKPDFView actionSheetAssistant:openElementAtLocation:]):
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::navigateToPDFLinkWithSimulatedClick): Renamed from navigateToURLWithSimulatedClick.
(WebKit::WebPage::navigateToURLWithSimulatedClick): Renamed to navigateToPDFLinkWithSimulatedClick.
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184441 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoREGRESSION (Subpixel): Dashed underline is missing when box is positioned at subpixels.
zalan@apple.com [Sat, 16 May 2015 23:24:00 +0000 (23:24 +0000)]
REGRESSION (Subpixel): Dashed underline is missing when box is positioned at subpixels.
https://bugs.webkit.org/show_bug.cgi?id=145097
rdar://problem/18588415

Reviewed by Simon Fraser.

Dashed and dotted border painting needs clipping in order to properly display corners.
Similarly to solid border's quad calculation, we pixelsnap the border positions before computing the clipping quad values.

Source/WebCore:

Test: fast/borders/dashed-border-on-subpixel-position.html
      fast/borders/dotted-border-on-subpixel-position.html

* rendering/RenderBoxModelObject.cpp:
(WebCore::RenderBoxModelObject::clipBorderSidePolygon):

LayoutTests:

* fast/borders/dashed-border-on-subpixel-position-expected.html: Added.
* fast/borders/dashed-border-on-subpixel-position.html: Added.
* fast/borders/dotted-border-on-subpixel-position-expected.html: Added.
* fast/borders/dotted-border-on-subpixel-position.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184440 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[iOS] Remote scrolling tree needs to coordinate scroll snap state during resize/rotations
bfulgham@apple.com [Sat, 16 May 2015 23:00:06 +0000 (23:00 +0000)]
[iOS] Remote scrolling tree needs to coordinate scroll snap state during resize/rotations
https://bugs.webkit.org/show_bug.cgi?id=145059
<rdar://problem/20975978>

Reviewed by Simon Fraser.

The web view needs to update its scroll snap point offsets to take into account any
adjustments to the view size caused by content insets. We also need to update the
offstes after device rotation.

* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView scrollViewWillEndDragging:withVelocity:targetContentOffset:]): Retrieve the proper
computed content inset for the view and incorporate into any scroll snap point adjustments.
(-[WKWebView _updateVisibleContentRects]): If we have active scroll snap points,
* UIProcess/Scrolling/RemoteScrollingCoordinatorProxy.h:
* UIProcess/ios/RemoteScrollingCoordinatorProxyIOS.mm:
(WebKit::RemoteScrollingCoordinatorProxy::adjustTargetContentOffsetForSnapping): Update to account
for content inset.
(WebKit::RemoteScrollingCoordinatorProxy::shouldSnapForMainFrameScrolling): Also validate that the
active index is valid.
(WebKit::RemoteScrollingCoordinatorProxy::closestSnapOffsetForMainFrameScrolling): Update to track
current active snap offset index.
(WebKit::RemoteScrollingCoordinatorProxy::hasActiveSnapPoint): Added.
(WebKit::RemoteScrollingCoordinatorProxy::nearestActiveSnapPoint): Added. It calculates the proper
scroll position incorporating any snap point and content insets.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184439 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed, rolling out r184415.
commit-queue@webkit.org [Sat, 16 May 2015 21:01:42 +0000 (21:01 +0000)]
Unreviewed, rolling out r184415.
https://bugs.webkit.org/show_bug.cgi?id=145096

Broke several tests (Requested by msaboff on #webkit).

Reverted changeset:

"Insert store barriers late so that IR transformations don't
have to worry about them"
https://bugs.webkit.org/show_bug.cgi?id=145015
http://trac.webkit.org/changeset/184415

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184438 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoRemove styleText() and applyStyleText() from InspectorStyle.
commit-queue@webkit.org [Sat, 16 May 2015 20:00:04 +0000 (20:00 +0000)]
Remove styleText() and applyStyleText() from InspectorStyle.
https://bugs.webkit.org/show_bug.cgi?id=145093

Patch by Sungmann Cho <sungmann.cho@navercorp.com> on 2015-05-16
Reviewed by Darin Adler.

styleText() and applyStyleText() in InspectorStyle are private aliases for the public
accessors in the same class but are rarely used. I think it would be better off removing
these aliases and updating all the current call sites to use the public accessors directly
as noted in InspectorStyleSheet.h.

No new tests, no behavior change.

* inspector/InspectorStyleSheet.cpp:
(WebCore::InspectorStyle::populateAllProperties):
* inspector/InspectorStyleSheet.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184437 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoRemove MessageType variant of addMessageToConsole() from ChromeClient.
commit-queue@webkit.org [Sat, 16 May 2015 19:57:26 +0000 (19:57 +0000)]
Remove MessageType variant of addMessageToConsole() from ChromeClient.
https://bugs.webkit.org/show_bug.cgi?id=145095

Patch by Sungmann Cho <sungmann.cho@navercorp.com> on 2015-05-16
Reviewed by Darin Adler.

All addMessageToConsole() in ChromeClient are exactly the same thing.
So there is no reason to keep all this methods.

No new tests, no behavior change.

* page/ChromeClient.h:
* page/PageConsoleClient.cpp:
(WebCore::PageConsoleClient::messageWithTypeAndLevel):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184436 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoAdd getElementById to DocumentFragment
weinig@apple.com [Sat, 16 May 2015 18:16:26 +0000 (18:16 +0000)]
Add getElementById to DocumentFragment
https://bugs.webkit.org/show_bug.cgi?id=145094

Reviewed by Alexey Proskuryakov.

Source/WebCore:

Match the latest DOM standard and other browsers by adding getElementById
to DocumentFragment in addition to Document and SVGSVGElement. Add NonElementParentNode
interface that the DOM standard defines.

Test: fast/dom/DocumentFragment/getElementById.html

* DerivedSources.make:
* WebCore.xcodeproj/project.pbxproj:
Add NonElementParentNode.idl

* dom/Document.idl:
Mark Document as implementing NonElementParentNode and remove now redundant getElementById
declaration.

* dom/DocumentFragment.cpp:
(WebCore::DocumentFragment::getElementById):
* dom/DocumentFragment.h:
Add implementation of getElementById for DocumentFragments. Add a fast path for ShadowRoots
which can take advantage of the fact that they are TreeScopes to use the elements by id cache
in TreeScrope.

* dom/DocumentFragment.idl:
Mark DocumentFragment as implementing NonElementParentNode.

* dom/NonElementParentNode.idl: Copied from Source/WebCore/dom/NonDocumentTypeChildNode.idl.
Added.

* svg/SVGSVGElement.cpp:
(WebCore::SVGSVGElement::getElementById):
* svg/SVGSVGElement.h:
* svg/SVGSVGElement.idl:
Convert to taking an AtomicString to match other getElementByIds.

LayoutTests:

* fast/dom/DocumentFragment/getElementById-expected.txt: Added.
* fast/dom/DocumentFragment/getElementById.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184435 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWhen redirecting to data URL use HTTP response for same origin policy checks
antti@apple.com [Sat, 16 May 2015 13:53:21 +0000 (13:53 +0000)]
When redirecting to data URL use HTTP response for same origin policy checks
https://bugs.webkit.org/show_bug.cgi?id=145054
rdar://problem/20299050

Reviewed by Alexey Proskuryakov.

Source/WebCore:

Test: http/tests/security/canvas-remote-read-data-url-image-redirect.html

* dom/ScriptElement.cpp:
(WebCore::ScriptElement::notifyFinished):
* dom/ScriptExecutionContext.cpp:
(WebCore::ScriptExecutionContext::sanitizeScriptError):
* html/canvas/CanvasRenderingContext.cpp:
(WebCore::CanvasRenderingContext::wouldTaintOrigin):
* loader/ImageLoader.cpp:
(WebCore::ImageLoader::notifyFinished):
* loader/MediaResourceLoader.cpp:
(WebCore::MediaResourceLoader::responseReceived):
* loader/TextTrackLoader.cpp:
(WebCore::TextTrackLoader::notifyFinished):
* loader/cache/CachedImage.cpp:
(WebCore::CachedImage::isOriginClean):
* loader/cache/CachedResource.cpp:
(WebCore::CachedResource::passesAccessControlCheck):
(WebCore::CachedResource::passesSameOriginPolicyCheck):

    Factor repeatedly used same origin policy test into a function.

(WebCore::CachedResource::redirectReceived):

    When redirecting to a data URL save the redirect response.

(WebCore::CachedResource::responseForSameOriginPolicyChecks):

    In case we got redirected to data use that response instead of the final data response for policy checks.

* loader/cache/CachedResource.h:

LayoutTests:

* http/tests/security/canvas-remote-read-data-url-image-redirect-expected.txt: Added.
* http/tests/security/canvas-remote-read-data-url-image-redirect.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184434 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[iOS] wireless playback picker button is drawn incorrectly
jonlee@apple.com [Sat, 16 May 2015 07:52:22 +0000 (07:52 +0000)]
[iOS] wireless playback picker button is drawn incorrectly
https://bugs.webkit.org/show_bug.cgi?id=145057
rdar://problem/20877518

Reviewed by Jer Noble.

* Modules/mediacontrols/mediaControlsiOS.js:
(ControllerIOS.prototype.updateWirelessTargetPickerButton): This method is
defined in Controller, but is only meant to run on Macs. Override it with
an empty function for iOS.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184433 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[EFL][WK2] Minibrowser : Add search icon to search bar
commit-queue@webkit.org [Sat, 16 May 2015 07:11:01 +0000 (07:11 +0000)]
[EFL][WK2] Minibrowser : Add search icon to search bar
https://bugs.webkit.org/show_bug.cgi?id=139773

Patch by Tanay C <tanay.c@samsung.com> on 2015-05-16
Reviewed by Gyuyoung Kim.

* MiniBrowser/efl/main.c:
(search_icon_show): Added.
(search_box_show): Modified.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184432 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agorun_benchmark should have an option to specify the number of runs
rniwa@webkit.org [Sat, 16 May 2015 06:42:45 +0000 (06:42 +0000)]
run_benchmark should have an option to specify the number of runs
https://bugs.webkit.org/show_bug.cgi?id=145091

Reviewed by Stephanie Lewis.

Added --count option.

* Scripts/run-benchmark:
(main):
* Scripts/webkitpy/benchmark_runner/benchmark_runner.py:
(BenchmarkRunner.__init__):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184431 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoREGRESSION (r183300): Background missing on top links on apple.com
simon.fraser@apple.com [Sat, 16 May 2015 04:24:48 +0000 (04:24 +0000)]
REGRESSION (r183300): Background missing on top links on apple.com
https://bugs.webkit.org/show_bug.cgi?id=145079
rdar://problem/20914252

Reviewed by Tim Horton.

Source/WebCore:

Re-land r184421 with a fix to check against rects.coverageRect
rather than the stale m_coverageRect.

Test: compositing/visible-rect/backing-change-height-from-zero.html

* platform/graphics/ca/GraphicsLayerCA.cpp:
(WebCore::GraphicsLayerCA::setVisibleAndCoverageRects):
* rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::shouldSkipLayerInDump):

LayoutTests:

Re-land.

* compositing/visible-rect/backing-change-height-from-zero-expected.txt: Added.
* compositing/visible-rect/backing-change-height-from-zero.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184430 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWeb Inspector: empty timeline should not use previous timeline's zoom interval
mattbaker@apple.com [Sat, 16 May 2015 03:47:09 +0000 (03:47 +0000)]
Web Inspector: empty timeline should not use previous timeline's zoom interval
https://bugs.webkit.org/show_bug.cgi?id=132754

Reviewed by Joseph Pecoraro.

When a timeline recording is started in response to a provisional load, TimelineManager should check whether
the main resource url is changing before loading a new timeline recording. If the main resource is changing,
set the selection start, selection duration, and duration-per-pixel settings of the TimelineOverview to their
default values.

* UserInterface/Controllers/TimelineManager.js:
(WebInspector.TimelineManager):
(WebInspector.TimelineManager.prototype.isCapturingPageReload):
(WebInspector.TimelineManager.prototype.capturingStopped):
(WebInspector.TimelineManager.prototype.pageDidLoad):
(WebInspector.TimelineManager.prototype._startAutoCapturing):
Added a property for checking whether auto capture was triggered by a page reload.

* UserInterface/Views/LinearTimelineOverview.js:
(WebInspector.LinearTimelineOverview):
Increase default selection time to 15 seconds.

* UserInterface/Views/TimelineOverview.js:
(WebInspector.TimelineOverview):
Reset selection and zoom if TimelineManager isn't capturing in response to a page reload.

(WebInspector.TimelineOverview.prototype.reset):
Reset selection and zoom.

(WebInspector.TimelineOverview.prototype._timeRangeSelectionChanged):
(WebInspector.TimelineOverview.prototype._resetSelection):
Added a helper function for resetting selection and zoom to their default values.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184429 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWeb Inspector: REGRESSION (r181625): Timeline recording started from console.profile...
mattbaker@apple.com [Sat, 16 May 2015 03:46:25 +0000 (03:46 +0000)]
Web Inspector: REGRESSION (r181625): Timeline recording started from console.profile is always empty
https://bugs.webkit.org/show_bug.cgi?id=144882

Reviewed by Joseph Pecoraro.

A timeline recording will always be stopped before the end of the current run loop is observed. Before
dispatching the recording stopped event, any events that are currently in progress should be considered
complete and sent to the frontend.

* inspector/InspectorTimelineAgent.cpp:
(WebCore::InspectorTimelineAgent::internalStop):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184428 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWeb Inspector: Timeline data grid displays wrong records after switching between...
mattbaker@apple.com [Sat, 16 May 2015 03:00:05 +0000 (03:00 +0000)]
Web Inspector: Timeline data grid displays wrong records after switching between Timelines/Frames mode
https://bugs.webkit.org/show_bug.cgi?id=145084

Reviewed by Timothy Hatcher.

Update timeline sidebar filter whenever the view mode changes.

* UserInterface/Views/TimelineSidebarPanel.js:
(WebInspector.TimelineSidebarPanel.prototype.treeElementForRepresentedObject.get if.get if):
(WebInspector.TimelineSidebarPanel.get else):
(WebInspector.TimelineSidebarPanel.prototype.treeElementForRepresentedObject.get if):
(WebInspector.TimelineSidebarPanel.prototype.treeElementForRepresentedObject):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184427 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWeb Inspector: Bump image format number to force image regeneration
commit-queue@webkit.org [Sat, 16 May 2015 02:31:34 +0000 (02:31 +0000)]
Web Inspector: Bump image format number to force image regeneration
https://bugs.webkit.org/show_bug.cgi?id=145074

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2015-05-15
Reviewed by Timothy Hatcher.

* UserInterface/Base/ImageUtilities.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184426 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoDon't reset the preview recognizer in [WKContentViewInteraction cleanupInteraction]
commit-queue@webkit.org [Sat, 16 May 2015 02:24:19 +0000 (02:24 +0000)]
Don't reset the preview recognizer in [WKContentViewInteraction cleanupInteraction]
https://bugs.webkit.org/show_bug.cgi?id=145081

We shouldn't reset the recognizer in [WKContentViewInteraction cleanupInteraction] since we don't re-add it
back to WKContentView when the process relaunched. Since we already add/remove it when the view is move into/
removed from the window, we don't need reset it in cleanupInteraction.

Patch by Yongjun Zhang <yongjun_zhang@apple.com> on 2015-05-15
Reviewed by Dan Bernstein.

* UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView cleanupInteraction]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184425 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoTest the interaction between font-family and font-weight
mmaxfield@apple.com [Sat, 16 May 2015 01:46:53 +0000 (01:46 +0000)]
Test the interaction between font-family and font-weight
https://bugs.webkit.org/show_bug.cgi?id=145078

Reviewed by Dean Jackson.

This test is designed to test the interaction between font-family and font-weight. In particular,
our implementation of font-family accepts PostScript names, which may name a font with a particular
weight. However, there is another CSS property, font-weight, in which the author may also name a
particular weight. Our font selection algorithm takes both of these signals into account when
choosing fonts.

There is currently no good way in JavaScript to find the actual font chosen for some text.
Therefore, the best way to test this aspect of the font selection algorithm is to dump the render
tree, therefore testing glyph advances (which are a property of font weight).

* platform/mac/fast/text/font-weights-expected.txt: Added.
* platform/mac/fast/text/font-weights.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184424 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed, rolling out r184421.
commit-queue@webkit.org [Sat, 16 May 2015 01:37:26 +0000 (01:37 +0000)]
Unreviewed, rolling out r184421.
https://bugs.webkit.org/show_bug.cgi?id=145087

Introduced 42 layout test failures (Requested by rniwa on
#webkit).

Reverted changeset:

"REGRESSION (r183300): Background missing on top links on
apple.com"
https://bugs.webkit.org/show_bug.cgi?id=145079
http://trac.webkit.org/changeset/184421

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184423 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoREGRESSION (r181910): WKWebView incorrectly scales snapshot
commit-queue@webkit.org [Sat, 16 May 2015 00:21:46 +0000 (00:21 +0000)]
REGRESSION (r181910): WKWebView incorrectly scales snapshot
https://bugs.webkit.org/show_bug.cgi?id=145076

Patch by James Savage <james.savage@apple.com> on 2015-05-15
Reviewed by Tim Horton.

We added a fast path to snapshotting using IOSurfaces with
http://trac.webkit.org/changeset/181910 which incorrectly determined
scale and transform and resulted in bad snapshots in some situations.

* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _snapshotRect:intoImageOfWidth:completionHandler:]):
When snapshotting an IOSurface we need to compute our scale based off of
the rect in the WKWebView coordinates, we also failed to account for
non-zero origins in the snapshot rect. Remove an unused varible while
we're here.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184422 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoREGRESSION (r183300): Background missing on top links on apple.com
simon.fraser@apple.com [Fri, 15 May 2015 23:29:29 +0000 (23:29 +0000)]
REGRESSION (r183300): Background missing on top links on apple.com
https://bugs.webkit.org/show_bug.cgi?id=145079
rdar://problem/20914252

Reviewed by Tim Horton.

Source/WebCore:

GraphicsLayerCA::setVisibleAndCoverageRects() only set the m_intersectsCoverageRect
flag if the coverage rect changed, but it doesn't if you simply change the size of
the layer.

Instead, always re-evaluate the intersection, and set the CoverageRectChanged bit
if it changes.

Test: compositing/visible-rect/backing-change-height-from-zero.html

* platform/graphics/ca/GraphicsLayerCA.cpp:
(WebCore::GraphicsLayerCA::setVisibleAndCoverageRects):
* rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::shouldSkipLayerInDump): Don't skip any layers in a
debug layer dump, to assist debugging.

LayoutTests:

Composited box that toggles to a non-zero height and dumps layers.

* compositing/visible-rect/backing-change-height-from-zero-expected.txt: Added.
* compositing/visible-rect/backing-change-height-from-zero.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184421 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoMove HTMLElement's children property to ParentNode
weinig@apple.com [Fri, 15 May 2015 22:55:53 +0000 (22:55 +0000)]
Move HTMLElement's children property to ParentNode
https://bugs.webkit.org/show_bug.cgi?id=145072

Reviewed by Chris Dumez.

Source/WebCore:

Match the latest DOM standard and other browsers by moving the children property
to ParentNode, thus exposing it on Element, Document and DocumentFragment.

Test: fast/dom/ParentNode-children.html

* dom/ContainerNode.cpp:
(WebCore::ContainerNode::children):
Moved implementation from HTMLElement to here.

(WebCore::ContainerNode::ensureCachedHTMLCollection):
(WebCore::ContainerNode::cachedHTMLCollection):
Moved implementation from Element to here so that ContainerNode::children
can use it.

(WebCore::ContainerNode::firstElementChild):
(WebCore::ContainerNode::lastElementChild):
(WebCore::ContainerNode::childElementCount):
Remove unnecessary assertions.

* dom/ContainerNode.h:
Moved declarations from HTMLElement and Element to here.

* dom/Element.cpp:
(WebCore::Element::ensureCachedHTMLCollection): Deleted.
(WebCore::Element::cachedHTMLCollection): Deleted.
* dom/Element.h:
Moved to ContainerNode.

* dom/Node.cpp:
(WebCore::Node::previousElementSibling):
(WebCore::Node::nextElementSibling):
Remove unnecessary assertions.

* dom/ParentNode.idl:
Moved the children property here from HTMLElement.

* html/HTMLElement.cpp:
(WebCore::HTMLElement::children): Deleted.
* html/HTMLElement.h:
Moved to ContainerNode.

* html/HTMLElement.idl:
Get rid of the children property for all bindings except Objective-C which needs
to keep it to avoid changing the public API.

LayoutTests:

* fast/dom/ParentNode-children-expected.txt: Added.
* fast/dom/ParentNode-children.html: Added.
Test children on Element, Document and DocumentFragment.

* fast/dom/xmlserializer-serialize-to-string-exception-expected.txt:
Update results since document.children is no longer undefined.

* js/dom/dom-static-property-for-in-iteration-expected.txt:
Update results.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184420 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWebCore ASan debug build fails. ERROR: WebCore has a weak external symbol in it.
zalan@apple.com [Fri, 15 May 2015 21:59:50 +0000 (21:59 +0000)]
WebCore ASan debug build fails. ERROR: WebCore has a weak external symbol in it.
https://bugs.webkit.org/show_bug.cgi?id=145070

Reviewed by David Kilzer.

Build fix.

* Configurations/WebCore.unexp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184419 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWeb Inspector: Reduce type annotation update frequency
commit-queue@webkit.org [Fri, 15 May 2015 21:36:02 +0000 (21:36 +0000)]
Web Inspector: Reduce type annotation update frequency
https://bugs.webkit.org/show_bug.cgi?id=145066

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2015-05-15
Reviewed by Timothy Hatcher.

* UserInterface/Controllers/TypeTokenAnnotator.js:
Reduce the frequency which could commonly be 16-24ms
to a minimum of 100ms and maximum of 2000ms.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184418 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoPrefix force on MouseEvent and add constants for click and force click values
bdakin@apple.com [Fri, 15 May 2015 21:31:25 +0000 (21:31 +0000)]
Prefix force on MouseEvent and add constants for click and force click values
https://bugs.webkit.org/show_bug.cgi?id=145065
-and corresponding-
rdar://problem/20770052

Reviewed by Tim Horton.

(WebCore::MouseEvent::webkitForce):
* dom/MouseEvent.idl:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184417 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[MediaControls] Refactor media controls & bring improvements made to iOS controls...
jer.noble@apple.com [Fri, 15 May 2015 21:30:18 +0000 (21:30 +0000)]
[MediaControls] Refactor media controls & bring improvements made to iOS controls to Mac.
https://bugs.webkit.org/show_bug.cgi?id=144973

Reviewed by Dean Jackson.

Pull improvements made to the iOS media controls back into the Mac controls by moving
code from mediaControlsiOS.js into MediaControlsApple.js.

The largest refactored feature is the ability to drop individual controls from the media
controls when the video is too small to contain them. To allow these controls to resize
dynamically, a new "resize" event is fired inside the media element's shadow DOM.

* Modules/mediacontrols/mediaControlsApple.css:
(audio::-webkit-media-controls-panel .dropped): Added; sets "display: none".
* Modules/mediacontrols/mediaControlsApple.js:
(Controller): Set defaults for new variables.
(Controller.prototype.updateControls): Update the controls width; moved from iOS.js.
(Controller.prototype.handleReadyStateChange): Update the controls; moved from iOS.js.
(Controller.prototype.handleTimeUpdate): Update the progress; moved from iOS.js.
(Controller.prototype.handleTimelineInput): Pause if scrubbing; moved from iOS.js.
(Controller.prototype.handleTimelineChange): Update the progress; moved from iOS.js.
(Controller.prototype.showControls): Update the controls width; moved from iOS.js.
(Controller.prototype.hideControls): Removed _potentiallyScrubbing check; not needed due to changes
    to controlsAlwaysVisible().
(Controller.prototype.scheduleUpdateLayoutForDisplayedWidth): Moved from iOS.js.
(Controller.prototype.isControlVisible): Added; checks whether control is parented & not hidden.
(Controller.prototype.updateLayoutForDisplayedWidth): Moved from iOS.js and refactored.
(Controller.prototype.controlsAlwaysVisible): Return true if scrubbing.
(Controller.prototype.updateHasAudio): Check currentPlaybackTargetIsWireless(); moved from iOS.js.
(Controller.prototype.get scrubbing): Simple getter for _scrubbing.
(Controller.prototype.set scrubbing): Check play state if scrubbing; start playback (if necessary)
    if not scrubbing.
(Controller.prototype.get pageScaleFactor): Moved from iOS.js.
(Controller.prototype.set pageScaleFactor): Ditto.
(Controller.prototype.handleRootResize): Schedule an update of the contrtols width.

Remove a bunch of newly unnecessary code from the iOS media controls:

* Modules/mediacontrols/mediaControlsiOS.js:
(ControllerIOS):
(ControllerIOS.prototype.createControls): Remove ivars moved into Apple.js.
(ControllerIOS.prototype.configureInlineControls): Remove spacer; made unnecessary.
(ControllerIOS.prototype.showControls): Deleted.
(ControllerIOS.prototype.updateTime): Deleted.
(ControllerIOS.prototype.handleTimelineTouchStart): Just call "scrubbing = true", handled in Apple.js.
(ControllerIOS.prototype.handleTimelineTouchEnd): Just call "scrubbing = false", handled in Apple.js.
(ControllerIOS.prototype.handleReadyStateChange): Deleted.
(ControllerIOS.prototype.setPlaying): Don't check _timelineIsHidden; not needed.
(ControllerIOS.prototype.get pageScaleFactor): Deleted.
(ControllerIOS.prototype.set pageScaleFactor): Deleted.
(ControllerIOS.prototype.scheduleUpdateLayoutForDisplayedWidth): Deleted.
(ControllerIOS.prototypeupdateLayoutForDisplayedWidth): Deleted.

Fire a "resize" event at the shadow DOM root when layout results in a size change.

* html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::layoutSizeChanged): Fire the "resize" event at the shadow DOM.
* html/HTMLMediaElement.h:
* rendering/RenderMedia.cpp:
(WebCore::RenderMedia::layout): Trigger layoutSizeChanged()
* rendering/RenderMedia.h:

Drive-by fixes:

* Modules/mediacontrols/mediaControlsApple.js:
(Controller.prototype.createControls): aria-label text is totally wrong; removed.
(Controller.prototype.updateWirelessPlaybackStatus): Use class-names to hide controls, not inline styles.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184416 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoInsert store barriers late so that IR transformations don't have to worry about them
fpizlo@apple.com [Fri, 15 May 2015 21:11:39 +0000 (21:11 +0000)]
Insert store barriers late so that IR transformations don't have to worry about them
https://bugs.webkit.org/show_bug.cgi?id=145015

Reviewed by Geoffrey Garen.

We have had three kinds of bugs with store barriers. For the sake of discussion we say
that a store barrier is needed when we have something like:

    base.field = value

- We sometimes fail to realize that we could remove a barrier when value is a non-cell.
  This might happen if we prove value to be a non-cell even though in the FixupPhase it
  wasn't predicted non-cell.

- We sometimes have a barrier in the wrong place after object allocation sinking. We
  might sink an allocation to just above the store, but that puts it just after the
  StoreBarrier that FixupPhase inserted.

- We don't remove redundant barriers across basic blocks.

This comprehensively fixes these issues by doing store barrier insertion late, and
removing the store barrier elision phase. Store barrier insertion uses an epoch-based
algorithm to determine when stores need barriers. Briefly, a barrier is not needed if
base is in the current GC epoch (i.e. was the last object that we allocated or had a
barrier since last GC) or if base has a newer GC epoch than value (i.e. value would have
always been allocated before base). We do conservative things when merging epoch state
between basic blocks, and we only do such inter-block removal in the FTL. FTL also
queries AI to determine what type we've proved about value, and avoids barriers when
value is not a cell. FixupPhase still inserts type checks on some stores, to maximize
the likelihood that this AI-based removal is effective.

* CMakeLists.txt:
* JavaScriptCore.vcxproj/JavaScriptCore.vcxproj:
* JavaScriptCore.xcodeproj/project.pbxproj:
* dfg/DFGBlockMap.h:
(JSC::DFG::BlockMap::at):
* dfg/DFGConstantFoldingPhase.cpp:
(JSC::DFG::ConstantFoldingPhase::emitPutByOffset):
* dfg/DFGEpoch.h:
(JSC::DFG::Epoch::operator<):
(JSC::DFG::Epoch::operator>):
(JSC::DFG::Epoch::operator<=):
(JSC::DFG::Epoch::operator>=):
* dfg/DFGFixupPhase.cpp:
(JSC::DFG::FixupPhase::fixupNode):
(JSC::DFG::FixupPhase::speculateForBarrier):
(JSC::DFG::FixupPhase::insertStoreBarrier): Deleted.
* dfg/DFGPlan.cpp:
(JSC::DFG::Plan::compileInThreadImpl):
* dfg/DFGStoreBarrierElisionPhase.cpp: Removed.
* dfg/DFGStoreBarrierElisionPhase.h: Removed.
* dfg/DFGStoreBarrierInsertionPhase.cpp: Added.
(JSC::DFG::performFastStoreBarrierInsertion):
(JSC::DFG::performGlobalStoreBarrierInsertion):
* dfg/DFGStoreBarrierInsertionPhase.h: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184415 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[ARM64] Do not fail branchConvertDoubleToInt32 when the result is zero and not negati...
benjamin@webkit.org [Fri, 15 May 2015 21:10:11 +0000 (21:10 +0000)]
[ARM64] Do not fail branchConvertDoubleToInt32 when the result is zero and not negative zero
https://bugs.webkit.org/show_bug.cgi?id=144976

Patch by Benjamin Poulain <bpoulain@apple.com> on 2015-05-15
Reviewed by Michael Saboff.

Failing the conversion on zero is pretty dangerous as we discovered on x86.

This patch does not really impact performance significantly because
r184220 removed the zero checks from Kraken. This patch is just to be
on the safe side for cases not covered by existing benchmarks.

* assembler/MacroAssemblerARM64.h:
(JSC::MacroAssemblerARM64::branchConvertDoubleToInt32):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184414 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[Content Extensions] Fail to load old content extension files
achristensen@apple.com [Fri, 15 May 2015 20:48:21 +0000 (20:48 +0000)]
[Content Extensions] Fail to load old content extension files
https://bugs.webkit.org/show_bug.cgi?id=145027

Reviewed by Eric Carlson.

Source/WebCore:

* contentextensions/DFABytecode.h:

Source/WebKit2:

* UIProcess/API/APIUserContentExtensionStore.cpp:
(API::openAndMapContentExtension):
* UIProcess/API/APIUserContentExtensionStore.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184413 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoRemove unnecessary forward declarations in PropertyNameArray.h.
commit-queue@webkit.org [Fri, 15 May 2015 20:47:16 +0000 (20:47 +0000)]
Remove unnecessary forward declarations in PropertyNameArray.h.
https://bugs.webkit.org/show_bug.cgi?id=145058

Patch by Sungmann Cho <sungmann.cho@navercorp.com> on 2015-05-15
Reviewed by Andreas Kling.

No new tests, no behavior change.

* runtime/PropertyNameArray.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184412 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWeb Inspector: Fix some possible event listener leakers in content views
commit-queue@webkit.org [Fri, 15 May 2015 20:47:08 +0000 (20:47 +0000)]
Web Inspector: Fix some possible event listener leakers in content views
https://bugs.webkit.org/show_bug.cgi?id=145068

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2015-05-15
Reviewed by Timothy Hatcher.

Now that some content views can be closed that weren't closable before, ensure
we clean up after global event listeners that may strongly reference "this".

* UserInterface/Views/DOMTreeContentView.js:
(WebInspector.DOMTreeContentView.prototype.closed):
* UserInterface/Views/DOMTreeOutline.js:
(WebInspector.DOMTreeOutline.prototype.close):
* UserInterface/Views/LayoutTimelineView.js:
(WebInspector.LayoutTimelineView.prototype.closed):
* UserInterface/Views/NetworkTimelineView.js:
(WebInspector.NetworkTimelineView.prototype.closed):
* UserInterface/Views/RenderingFrameTimelineView.js:
(WebInspector.RenderingFrameTimelineView.prototype.closed):
* UserInterface/Views/ScriptContentView.js:
(WebInspector.ScriptContentView.prototype.closed):
* UserInterface/Views/ScriptTimelineView.js:
(WebInspector.ScriptTimelineView.prototype.closed):
* UserInterface/Views/SourceCodeTextEditor.js:
(WebInspector.SourceCodeTextEditor.prototype.close):
* UserInterface/Views/TextResourceContentView.js:
(WebInspector.TextResourceContentView.prototype.closed):
* UserInterface/Views/TimelineDataGrid.js:
(WebInspector.TimelineDataGrid.prototype.closed):
(WebInspector.TimelineDataGrid.prototype.handleEvent):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184411 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWeb Inspector: Initiator Popovers no longer work in Layout Timeline
commit-queue@webkit.org [Fri, 15 May 2015 20:45:58 +0000 (20:45 +0000)]
Web Inspector: Initiator Popovers no longer work in Layout Timeline
https://bugs.webkit.org/show_bug.cgi?id=145067

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2015-05-15
Reviewed by Timothy Hatcher.

* UserInterface/Views/LayoutTimelineDataGrid.js:
(WebInspector.LayoutTimelineDataGrid.prototype.callFramePopoverAnchorElement):
Update the column name, which changed in r183134.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184410 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoTemporarily disable failing API test.
timothy_horton@apple.com [Fri, 15 May 2015 20:45:17 +0000 (20:45 +0000)]
Temporarily disable failing API test.

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

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184409 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoJSArray::setLength() should reallocate instead of zero-filling if the reallocation...
mark.lam@apple.com [Fri, 15 May 2015 20:02:26 +0000 (20:02 +0000)]
JSArray::setLength() should reallocate instead of zero-filling if the reallocation would be small enough.
https://bugs.webkit.org/show_bug.cgi?id=144622

Reviewed by Geoffrey Garen.

When setting the array to a new length that is shorter, we now check if it is worth
just making a new butterfly instead of clearing out the slots in the old butterfly
that resides beyond the new length.  If so, we will make a new butterfly instead.

There is no perf differences in the benchmark results.  However, this does benefit
the perf of pathological cases where we need to shorten the length of a very large
array, as is the case in tests/mozilla/js1_5/Array/regress-101964.js.  With this
patch, we can expect that test to complete in a short time again.

* runtime/JSArray.cpp:
(JSC::JSArray::setLength):
* runtime/JSObject.cpp:
(JSC::JSObject::reallocateAndShrinkButterfly):
- makes a new butterfly with a new shorter length.
* runtime/JSObject.h:
* tests/mozilla/js1_5/Array/regress-101964.js:
- Undo this test change since this patch will prevent us from spending a lot of time
  clearing a large butterfly.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184407 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed build fix.
eric.carlson@apple.com [Fri, 15 May 2015 20:02:13 +0000 (20:02 +0000)]
Unreviewed build fix.

* platform/ios/WebVideoFullscreenInterfaceAVKit.mm: Add clang pragmas to ignore
deprecation warnings.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184406 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoDFGLICMPhase shouldn't create NodeOrigins with forExit but without semantic
basile_clement@apple.com [Fri, 15 May 2015 19:30:14 +0000 (19:30 +0000)]
DFGLICMPhase shouldn't create NodeOrigins with forExit but without semantic
https://bugs.webkit.org/show_bug.cgi?id=145062

Reviewed by Filip Pizlo.

We assert in various places (including NodeOrigin::isSet()) that a
NodeOrigin's semantic and forExit must be either both set, or both
unset.  However, LICM'ing a node with unset NodeOrigin would only set
forExit, and leave semantic unset. This can for instance happen when a
Phi node is constant-folded into a JSConstant, which in turn gets
LICM'd.

This patch changes DFGLICMPhase to set the NodeOrigin's semantic in
addition to its forExit if semantic was previously unset.

It also adds two validators to DFGValidate.cpp:
 - In both SSA and CPS form, a NodeOrigin semantic and forExit must be either both set or both unset
 - In CPS form, all nodes must have a set NodeOrigin forExit (this is
   the CPS counterpart to the SSA validator that checks that all nodes
   must have a set NodeOrigin except possibly for a continuous chunk of
   nodes at the top of a block)

* dfg/DFGLICMPhase.cpp:
(JSC::DFG::LICMPhase::attemptHoist):
* dfg/DFGValidate.cpp:
(JSC::DFG::Validate::validate):
(JSC::DFG::Validate::validateCPS):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184405 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoLimit alternate fullscreen with linked on or after.
commit-queue@webkit.org [Fri, 15 May 2015 19:12:58 +0000 (19:12 +0000)]
Limit alternate fullscreen with linked on or after.
https://bugs.webkit.org/show_bug.cgi?id=144894

Patch by Jeremy Jones <jeremyj@apple.com> on 2015-05-15
Reviewed by Dean Jackson.

Source/WebCore:

* platform/ios/WebCoreSystemInterfaceIOS.h: add new wkIOSSystemVersion

Source/WebKit/mac:

* WebView/WebView.mm:
(shouldAllowAlternateFullscreen): Added.
(-[WebView _preferencesChanged:]):

Source/WebKit2:

* UIProcess/API/Cocoa/WKWebView.mm:
(shouldAllowAlternateFullscreen): Added.
(-[WKWebView initWithFrame:configuration:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184404 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed. Revert part of r184361.
roger_fong@apple.com [Fri, 15 May 2015 19:07:28 +0000 (19:07 +0000)]
Unreviewed. Revert part of r184361.

* Modules/mediacontrols/mediaControlsApple.js:
(Controller.prototype.configureInlineControls):
HI wants the buttons flipped.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184403 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoFix trivial typo in TextEncodingDetectorICU.cpp.
commit-queue@webkit.org [Fri, 15 May 2015 18:44:12 +0000 (18:44 +0000)]
Fix trivial typo in TextEncodingDetectorICU.cpp.
https://bugs.webkit.org/show_bug.cgi?id=145055

Patch by Sungmann Cho <sungmann.cho@navercorp.com> on 2015-05-15
Reviewed by Alexey Proskuryakov.

No new tests, no behavior change..

* platform/text/TextEncodingDetectorICU.cpp:
(WebCore::detectTextEncoding):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184402 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoCyrillic top-level domains are displayed as punycode
ap@apple.com [Fri, 15 May 2015 18:41:54 +0000 (18:41 +0000)]
Cyrillic top-level domains are displayed as punycode
https://bugs.webkit.org/show_bug.cgi?id=145024
rdar://problem/17747133
rdar://problem/14116594

Reviewed by Tim Horton.

Handling each TLD in code is annoying, but we can probably survive like this
for a few more years, and maybe we'll think of an entirely different way to deal
with non-ASCII domain labels in the meanwhile.

* platform/mac/WebCoreNSURLExtras.mm:
(WebCore::isSecondLevelDomainNameAllowedByTLDRules):
(WebCore::allCharactersAllowedByTLDRules):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184401 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoCursor is displayed after full screen video controls fade away.
roger_fong@apple.com [Fri, 15 May 2015 18:40:09 +0000 (18:40 +0000)]
Cursor is displayed after full screen video controls fade away.
https://bugs.webkit.org/show_bug.cgi?id=145034.
<rdar://problem/20458604>

Reviewed by Jer Noble.

* Modules/mediacontrols/mediaControlsApple.css:
(video::-webkit-media-controls-panel):
Unnecessary cursor style is on the control panel while hidden but
the cursor will already be hidden anyways from being auto-hidden.
Causes style to change when controls are hidden,
which causes the cursor to reappear.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184400 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[WK2][Cocoa] Back swipe tab snapshot takes a long time to be removed on bing.com
cdumez@apple.com [Fri, 15 May 2015 18:25:57 +0000 (18:25 +0000)]
[WK2][Cocoa] Back swipe tab snapshot takes a long time to be removed on bing.com
https://bugs.webkit.org/show_bug.cgi?id=145061
<rdar://problem/20939743>

Reviewed by Tim Horton.

When swiping back from a video search result to the list of video
search results on bing.com, the back swipe gesture tab snapshot takes
~3 seconds to be removed, even though the page gets loaded almost
instantly from PageCache. The tab snapshot should be removed as soon as
the load is done.

The issue is that we only cleared the back swipe gesture tab snapshot
after PageClient::didFinishLoadForMainFrame() has been called. However,
PageClient::didFinishLoadForMainFrame() was only being called by
WebPageProxy if the main frame loaded *without* error. In case the main
frame loaded with an error, only WebPageProxy::didFailLoadForFrame() is
called, not WebPageProxy::didFinishLoadForFrame() and we would fail to
remove the gesture snapshot until the 3 seconds timeout.

This patch calls PageClient::didFinishLoadForMainFrame() from
WebPageProxy::didFailLoadForFrame() so we remove the snapshot in the
error case as well.

The reason didFailLoadForFrame() is being called on bing.com video
search results is because a "ping" load is aborted when the page is
entering PageCache. Aborting any kind of resource load sets a
"cancellation" error on the main document.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::didFailLoadForFrame):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184399 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed, remove an unused declaration.
fpizlo@apple.com [Fri, 15 May 2015 18:21:29 +0000 (18:21 +0000)]
Unreviewed, remove an unused declaration.

* dfg/DFGSpeculativeJIT.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184398 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoRemove unused constant-base and constant-value store barrier code in the DFG
fpizlo@apple.com [Fri, 15 May 2015 18:09:07 +0000 (18:09 +0000)]
Remove unused constant-base and constant-value store barrier code in the DFG
https://bugs.webkit.org/show_bug.cgi?id=145039

Reviewed by Andreas Kling.

Just killing dead code.

* dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::storeToWriteBarrierBuffer): Deleted.
(JSC::DFG::SpeculativeJIT::writeBarrier): Deleted.
* dfg/DFGSpeculativeJIT.h:
* dfg/DFGSpeculativeJIT32_64.cpp:
(JSC::DFG::SpeculativeJIT::writeBarrier):
* dfg/DFGSpeculativeJIT64.cpp:
(JSC::DFG::SpeculativeJIT::writeBarrier):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184397 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoCheckmark on OFF option of captions sometimes does not appear.
roger_fong@apple.com [Fri, 15 May 2015 17:58:26 +0000 (17:58 +0000)]
Checkmark on OFF option of captions sometimes does not appear.
https://bugs.webkit.org/show_bug.cgi?id=145060.
<rdar://problem/19388333>

Reviewed by Eric Carlson.

* Modules/mediacontrols/mediaControlsApple.js:
(Controller.prototype.buildCaptionMenu):
Audio tracks don’t have an OFF option.
Remove related code to allow OFF option for subtitles to work properly.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184396 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWhite edge on animating panel on http://rokkosunnyvale.com
zalan@apple.com [Fri, 15 May 2015 17:39:05 +0000 (17:39 +0000)]
White edge on animating panel on rokkosunnyvale.com
https://bugs.webkit.org/show_bug.cgi?id=144986
rdar://problem/20907683

Reviewed by Simon Fraser.

Background image geometry calculation needs to be based on the final painting size of the container
in order to accurately compute tile sizes, repeating positions etc.
The container's size is pixelsnapped at painting using absolute coordinates. This patch
ensures that we snap to the same size while computing background geometry.

Source/WebCore:

Test: fast/images/background-image-size-changes-fractional-position.html

* rendering/InlineFlowBox.cpp:
(WebCore::InlineFlowBox::paintBoxDecorations):
* rendering/RenderBox.cpp:
(WebCore::RenderBox::paintBoxDecorations):
(WebCore::RenderBox::paintBackground):
(WebCore::RenderBox::getBackgroundPaintedExtent):
(WebCore::RenderBox::computeBackgroundIsKnownToBeObscured):
(WebCore::RenderBox::maskClipRect):
(WebCore::RenderBox::repaintLayerRectsForImage): unable to get absolute coords.
* rendering/RenderBox.h:
* rendering/RenderBoxModelObject.cpp:
(WebCore::RenderBoxModelObject::paintFillLayerExtended):
(WebCore::RenderBoxModelObject::calculateBackgroundImageGeometry):
(WebCore::RenderBoxModelObject::getGeometryForBackgroundImage):
(WebCore::RenderBoxModelObject::boxShadowShouldBeAppliedToBackground):
* rendering/RenderBoxModelObject.h:
* rendering/RenderFieldset.cpp:
(WebCore::RenderFieldset::paintBoxDecorations):
* rendering/RenderImage.cpp:
(WebCore::RenderImage::boxShadowShouldBeAppliedToBackground):
(WebCore::RenderImage::computeBackgroundIsKnownToBeObscured):
* rendering/RenderImage.h:
* rendering/RenderLayer.cpp: unable to get absolute coords.
(WebCore::RenderLayer::calculateClipRects):
* rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::updateDirectlyCompositedBackgroundImage): currently not used.
* rendering/RenderObject.h:
(WebCore::RenderObject::computeBackgroundIsKnownToBeObscured):
(WebCore::RenderObject::backgroundIsKnownToBeObscured):
* rendering/RenderTable.cpp:
(WebCore::RenderTable::paintBoxDecorations):
* rendering/RenderTableCell.cpp:
(WebCore::RenderTableCell::boxShadowShouldBeAppliedToBackground):
* rendering/RenderTableCell.h:

LayoutTests:

* fast/backgrounds/hidpi-bitmap-background-repeat-on-subpixel-position-expected.html: progression.
* fast/images/background-image-size-changes-fractional-position-expected.html: Added.
* fast/images/background-image-size-changes-fractional-position.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184395 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoCrash in RenderFlowThread::popFlowThreadLayoutState() due to mismatched push/pop...
jer.noble@apple.com [Fri, 15 May 2015 17:15:16 +0000 (17:15 +0000)]
Crash in RenderFlowThread::popFlowThreadLayoutState() due to mismatched push/pop count
https://bugs.webkit.org/show_bug.cgi?id=145042

Reviewed by David Hyatt.

RenderFlowThread previously used a ListHashSet to store its stack of active objects. This
is problematic because, if the same object is pushed twice, only a single entry of that
object is added to the stack. After this occurs, a matching number of pushes will pop too
many items off the stack, causing a crash when popping a stack with zero items. This
specifically happens in FrameView::layout(), which will push its root renderer on the stack
of active items, and then ask the root to layout(), which will attempt to push itself on the
stack of active items.

Instead of a ListHashSet, use a Vector, which has similar memory characteristics and no
uniqueness requirements.

* rendering/RenderFlowThread.cpp:
(WebCore::RenderFlowThread::pushFlowThreadLayoutState):
(WebCore::RenderFlowThread::popFlowThreadLayoutState):
* rendering/RenderFlowThread.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184394 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoBuild fix for some versions of clang.
mitz@apple.com [Fri, 15 May 2015 16:46:46 +0000 (16:46 +0000)]
Build fix for some versions of clang.

* wtf/SaturatedArithmetic.h:
(signedAddOverflows):
(signedSubtractOverflows):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184393 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[GTK] Gardening 15th May.
commit-queue@webkit.org [Fri, 15 May 2015 15:32:11 +0000 (15:32 +0000)]
[GTK] Gardening 15th May.
https://bugs.webkit.org/show_bug.cgi?id=145047

Unreviewed.

Patch by Marcos Chavarría Teijeiro <mchavarria@igalia.com> on 2015-05-15

* platform/gtk/TestExpectations:
* platform/gtk/fast/events/ghostly-mousemoves-in-subframe-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184392 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[Curl] WebSocket platform part is not implemented.
peavo@outlook.com [Fri, 15 May 2015 13:50:05 +0000 (13:50 +0000)]
[Curl] WebSocket platform part is not implemented.
https://bugs.webkit.org/show_bug.cgi?id=144628

Reviewed by Darin Adler.

Add Curl platform code implementation for WebSockets.

* platform/network/curl/SocketStreamHandle.h:
(WebCore::SocketStreamHandle::create):
(WebCore::SocketStreamHandle::SocketData::SocketData):
* platform/network/curl/SocketStreamHandleCurl.cpp:
(WebCore::SocketStreamHandle::SocketStreamHandle):
(WebCore::SocketStreamHandle::~SocketStreamHandle):
(WebCore::SocketStreamHandle::platformSend):
(WebCore::SocketStreamHandle::platformClose):
(WebCore::SocketStreamHandle::readData):
(WebCore::SocketStreamHandle::sendData):
(WebCore::SocketStreamHandle::waitForAvailableData):
(WebCore::SocketStreamHandle::startThread):
(WebCore::SocketStreamHandle::stopThread):
(WebCore::SocketStreamHandle::didReceiveData):
(WebCore::SocketStreamHandle::didOpenSocket):
(WebCore::SocketStreamHandle::createCopy):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184389 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUser interruption while running of run-webkit-tests should also generate results...
ossy@webkit.org [Fri, 15 May 2015 12:22:12 +0000 (12:22 +0000)]
User interruption while running of run-webkit-tests should also generate results.html
https://bugs.webkit.org/show_bug.cgi?id=122154

Patch by Ravi Phaneendra Kasibhatla <r.kasibhatla@samsung.com> on 2015-05-15
Reviewed by Csaba Osztrogonác.

Generation of results.html on execution of run-webkit-tests happens only
on completion of entire layout tests run. It should be created even when
the execution has been interrupted - either by user (by pressing Ctrl+C)
or because of other interruptions (like exit-after-n-failures option).

* Scripts/webkitpy/layout_tests/controllers/layout_test_runner.py:
(LayoutTestRunner.run_tests):
* Scripts/webkitpy/layout_tests/controllers/manager.py:
(Manager.run):
* Scripts/webkitpy/layout_tests/models/test_run_results.py:
(TestRunResults.__init__):
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
(main):
* Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
(RunTest.test_keyboard_interrupt):
(MainTest.test_exception_handling):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184382 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[buildbot] Fix the URL of the performance bots
ossy@webkit.org [Fri, 15 May 2015 11:00:15 +0000 (11:00 +0000)]
[buildbot] Fix the URL of the performance bots
https://bugs.webkit.org/show_bug.cgi?id=145043

Reviewed by Ryosuke Niwa.

* BuildSlaveSupport/build.webkit.org-config/templates/root.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184379 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoFix typo in function name parseFunctionParamters -> parseFunctionParameters
commit-queue@webkit.org [Fri, 15 May 2015 09:22:19 +0000 (09:22 +0000)]
Fix typo in function name parseFunctionParamters -> parseFunctionParameters
https://bugs.webkit.org/show_bug.cgi?id=145040

Patch by Alexandr Skachkov <gskachkov@gmail.com> on 2015-05-15
Reviewed by Mark Lam.

* parser/Parser.h:
* parser/Parser.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184378 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoREGRESSION(r183861): [SOUP] Downloads are broken when using the Network Process
carlosgc@webkit.org [Fri, 15 May 2015 08:03:16 +0000 (08:03 +0000)]
REGRESSION(r183861): [SOUP] Downloads are broken when using the Network Process
https://bugs.webkit.org/show_bug.cgi?id=144738

Reviewed by Alexey Proskuryakov.

Source/WebCore:

Add ResourceHandle::releaseForDownload() that releases the current
handle to be used as a download.

* platform/network/ResourceHandle.h:
* platform/network/soup/ResourceHandleSoup.cpp:
(WebCore::ResourceHandle::releaseForDownload):

Source/WebKit2:

When converting the main resource handle to a download, the
NetworkResourceLoader is aborted, and the ResourceHandle is
cleaned up aborting the download operation. We need to use a
different ResourceHandle for the download operation.

* Shared/Downloads/soup/DownloadSoup.cpp:
(WebKit::Download::startWithHandle): Use ResourceHandle::releaseForDownload()
instead of reusing the given handle.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184376 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoRemoved failing test expectations from passing tests.
rniwa@webkit.org [Fri, 15 May 2015 08:01:42 +0000 (08:01 +0000)]
Removed failing test expectations from passing tests.

* TestExpectations:
* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184375 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[GTK] Fix PlatformDisplayWayland construction error, implement the destructor
zandobersek@gmail.com [Fri, 15 May 2015 07:05:47 +0000 (07:05 +0000)]
[GTK] Fix PlatformDisplayWayland construction error, implement the destructor
https://bugs.webkit.org/show_bug.cgi?id=144997

Reviewed by Carlos Garcia Campos.

The PlatformDisplayWayland constructor is private, so we can't use
std::make_unique<>() to construct an object of this class.

Implement the PlatformDisplayWayland destructor, cleaning out all
the Wayland resources, if present.

* platform/graphics/wayland/PlatformDisplayWayland.cpp:
(WebCore::PlatformDisplayWayland::create):
(WebCore::PlatformDisplayWayland::PlatformDisplayWayland):
(WebCore::PlatformDisplayWayland::~PlatformDisplayWayland):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184374 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoImages on www.fitstylelife.com jiggle on hover.
zalan@apple.com [Fri, 15 May 2015 05:09:46 +0000 (05:09 +0000)]
Images on www.fitstylelife.com jiggle on hover.
https://bugs.webkit.org/show_bug.cgi?id=145020
rdar://problem/20885337

Reviewed by Simon Fraser.

This patch ensures that the clipping layer of a composited content is pixel snapped properly.

Source/WebCore:

Tests: compositing/composited-parent-clipping-layer-on-subpixel-position.html
       compositing/parent-clipping-layer-on-subpixel-position.html

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

LayoutTests:

* compositing/composited-parent-clipping-layer-on-subpixel-position-expected.html: Added.
* compositing/composited-parent-clipping-layer-on-subpixel-position.html: Added.
* compositing/parent-clipping-layer-on-subpixel-position-expected.html: Added.
* compositing/parent-clipping-layer-on-subpixel-position.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184373 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoHave DOMWindow::createWindow() take references to frames
cdumez@apple.com [Fri, 15 May 2015 05:07:22 +0000 (05:07 +0000)]
Have DOMWindow::createWindow() take references to frames
https://bugs.webkit.org/show_bug.cgi?id=145037

Reviewed by Gyuyoung Kim.

Have DOMWindow::createWindow() take references to frames instead of
pointers as they are expected to be non-null. Also return a RefPtr
instead of a PassRefPtr.

* inspector/InspectorFrontendClientLocal.cpp:
(WebCore::InspectorFrontendClientLocal::openInNewTab):
* loader/FrameLoader.cpp:
(WebCore::createWindow):
* loader/FrameLoader.h:
* page/DOMWindow.cpp:
(WebCore::DOMWindow::createWindow):
(WebCore::DOMWindow::open):
(WebCore::DOMWindow::showModalDialog):
* page/DOMWindow.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184372 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoREGRESSION (r183794): Garbage tiles when body background switches to fixed
simon.fraser@apple.com [Fri, 15 May 2015 04:43:05 +0000 (04:43 +0000)]
REGRESSION (r183794): Garbage tiles when body background switches to fixed
https://bugs.webkit.org/show_bug.cgi?id=145032
rdar://problem/20963679

Reviewed by Dean Jackson.

Source/WebCore:

After r183794 (or possibly an earlier commit), we failed to dynamically update
the configuration of layers that handled fixed background attachment on the root.

This would result in unpainted tiles, and non-fixed-background behavior.

Fix by calling RenderLayerCompositor::rootOrBodyStyleChanged() whenever the
style changes on the root or body renderers, and triggering a compositing update
if the fixedness of the background changes. It calls the existing rootBackgroundTransparencyChanged()
if the color changes.

Test: platform/mac-wk2/tiled-drawing/toggle-to-fixed-background.html

* rendering/RenderBox.cpp:
(WebCore::RenderBox::styleDidChange):
* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::rootOrBodyStyleChanged):
(WebCore::RenderLayerCompositor::rootBackgroundTransparencyChanged):
* rendering/RenderLayerCompositor.h:

LayoutTests:

Test that toggles the attachment of the body background to fixed, then dumps layers.

* platform/mac-wk2/tiled-drawing/toggle-to-fixed-background-expected.txt: Added.
* platform/mac-wk2/tiled-drawing/toggle-to-fixed-background.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184371 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoRename connectionDidClose and related methods to be more clear.
beidson@apple.com [Fri, 15 May 2015 04:39:51 +0000 (04:39 +0000)]
Rename connectionDidClose and related methods to be more clear.
https://bugs.webkit.org/show_bug.cgi?id=145030

Reviewed by Darin Adler.

These methods were easy to confuse with "Connection::Client::didClose()", yet they
were about something much more explicit: A child process being shut down by the UI Process.

Let's call them as such.

* Shared/ChildProcessProxy.cpp:
(WebKit::ChildProcessProxy::shutDownProcess):
(WebKit::ChildProcessProxy::clearConnection): Deleted.
(WebKit::ChildProcessProxy::connectionDidClose): Deleted.
* Shared/ChildProcessProxy.h:

* UIProcess/Databases/DatabaseProcessProxy.cpp:
(WebKit::DatabaseProcessProxy::processWillShutDown):
* UIProcess/Databases/DatabaseProcessProxy.h:

* UIProcess/Network/NetworkProcessProxy.cpp:
(WebKit::NetworkProcessProxy::processWillShutDown):
* UIProcess/Network/NetworkProcessProxy.h:

* UIProcess/Plugins/PluginProcessProxy.cpp:
(WebKit::PluginProcessProxy::processWillShutDown):
* UIProcess/Plugins/PluginProcessProxy.h:

* UIProcess/WebFrameProxy.cpp:
(WebKit::WebFrameProxy::webProcessWillShutDown):
(WebKit::WebFrameProxy::disconnect): Deleted.
* UIProcess/WebFrameProxy.h:

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::webProcessWillShutDown):
(WebKit::WebPageProxy::connectionDidClose): Deleted.
* UIProcess/WebPageProxy.h:

* UIProcess/WebProcessLifetimeTracker.cpp:
(WebKit::WebProcessLifetimeTracker::webProcessWillShutDown):
(WebKit::WebProcessLifetimeTracker::connectionDidClose): Deleted.
* UIProcess/WebProcessLifetimeTracker.h:

* UIProcess/WebProcessProxy.cpp:
(WebKit::WebProcessProxy::processWillShutDown):
(WebKit::WebProcessProxy::shutDown):
(WebKit::WebProcessProxy::removeWebPage):
(WebKit::WebProcessProxy::didClose):
(WebKit::WebProcessProxy::disconnectFramesFromPage):
(WebKit::WebProcessProxy::shouldTerminate):
(WebKit::WebProcessProxy::requestTermination):
(WebKit::WebProcessProxy::connectionDidClose): Deleted.
(WebKit::WebProcessProxy::disconnect): Deleted.
* UIProcess/WebProcessProxy.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184370 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoReverted r177753, now that <rdar://problem/19347133> is fixed.
mitz@apple.com [Fri, 15 May 2015 04:36:27 +0000 (04:36 +0000)]
Reverted r177753, now that <rdar://problem/19347133> is fixed.

Rubber-stamped by Benjamin Poulain.

* wtf/SaturatedArithmetic.h:
(signedAddOverflows):
(signedSubtractOverflows):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184369 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoRemove StoreBarrierWithNullCheck, nobody ever generates this.
fpizlo@apple.com [Fri, 15 May 2015 04:14:39 +0000 (04:14 +0000)]
Remove StoreBarrierWithNullCheck, nobody ever generates this.

Rubber stamped by Benjamin Poulain and Michael Saboff.

If we did bring something like this back in the future, we would just use UntypedUse instead
of CellUse to indicate that this is what we want.

* dfg/DFGAbstractInterpreterInlines.h:
(JSC::DFG::AbstractInterpreter<AbstractStateType>::executeEffects):
* dfg/DFGClobberize.h:
(JSC::DFG::clobberize):
* dfg/DFGDoesGC.cpp:
(JSC::DFG::doesGC):
* dfg/DFGFixupPhase.cpp:
(JSC::DFG::FixupPhase::fixupNode):
* dfg/DFGNode.h:
(JSC::DFG::Node::isStoreBarrier):
* dfg/DFGNodeType.h:
* dfg/DFGObjectAllocationSinkingPhase.cpp:
(JSC::DFG::ObjectAllocationSinkingPhase::lowerNonReadingOperationsOnPhantomAllocations):
(JSC::DFG::ObjectAllocationSinkingPhase::handleNode):
* dfg/DFGPredictionPropagationPhase.cpp:
(JSC::DFG::PredictionPropagationPhase::propagate):
* dfg/DFGSafeToExecute.h:
(JSC::DFG::safeToExecute):
* dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::compileStoreBarrier):
* dfg/DFGSpeculativeJIT32_64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
* dfg/DFGSpeculativeJIT64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
* ftl/FTLCapabilities.cpp:
(JSC::FTL::canCompile):
* ftl/FTLLowerDFGToLLVM.cpp:
(JSC::FTL::LowerDFGToLLVM::compileNode):
(JSC::FTL::LowerDFGToLLVM::compileStoreBarrierWithNullCheck): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184368 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoPutGlobalVar should reference the global object it's storing into
fpizlo@apple.com [Fri, 15 May 2015 03:51:52 +0000 (03:51 +0000)]
PutGlobalVar should reference the global object it's storing into
https://bugs.webkit.org/show_bug.cgi?id=145036

Reviewed by Michael Saboff.

This makes it easier to reason about store barrier insertion and elimination. This changes
the format of PutGlobalVar so that child1 is the global object and child2 is the value.
Previously it just had child1, and that was the value.

* dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::parseBlock):
* dfg/DFGClobberize.h:
(JSC::DFG::clobberize):
* dfg/DFGFixupPhase.cpp:
(JSC::DFG::FixupPhase::fixupNode):
* dfg/DFGSpeculativeJIT32_64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
* dfg/DFGSpeculativeJIT64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
* ftl/FTLLowerDFGToLLVM.cpp:
(JSC::FTL::LowerDFGToLLVM::compilePutGlobalVar):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184367 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed, rolling out r184359 and r184362.
commit-queue@webkit.org [Fri, 15 May 2015 03:28:24 +0000 (03:28 +0000)]
Unreviewed, rolling out r184359 and r184362.
https://bugs.webkit.org/show_bug.cgi?id=145035

Introduced a crash in six media element tests (Requested by
rniwa on #webkit).

Reverted changesets:

"[MediaControls] Refactor media controls & bring improvements
made to iOS controls to Mac."
https://bugs.webkit.org/show_bug.cgi?id=144973
http://trac.webkit.org/changeset/184359

"Unreviewed build fix after r184359; typo."
http://trac.webkit.org/changeset/184362

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184366 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoSome CFNetwork SPI to reset HSTS hosts added since a date should not be used on Yosemite.
commit-queue@webkit.org [Fri, 15 May 2015 02:03:23 +0000 (02:03 +0000)]
Some CFNetwork SPI to reset HSTS hosts added since a date should not be used on Yosemite.
https://bugs.webkit.org/show_bug.cgi?id=145025.
and
rdar://problem/20646308.

Patch by Zhuo Li <zachli@apple.com> on 2015-05-14
Reviewed by Alexey Proskuryakov.

* UIProcess/Cocoa/WebProcessPoolCocoa.mm:
(WebKit::WebProcessPool::resetHSTSHostsAddedAfterDate):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184365 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWeb Inspector: Update the New Tab button disabled state after extra domains are activated
commit-queue@webkit.org [Fri, 15 May 2015 00:25:07 +0000 (00:25 +0000)]
Web Inspector: Update the New Tab button disabled state after extra domains are activated
https://bugs.webkit.org/show_bug.cgi?id=145028

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2015-05-14
Reviewed by Timothy Hatcher.

* UserInterface/Base/Main.js:
(WebInspector.activateExtraDomains):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184364 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[CMake] Error out when ruby is too old
mcatanzaro@igalia.com [Fri, 15 May 2015 00:08:51 +0000 (00:08 +0000)]
[CMake] Error out when ruby is too old
https://bugs.webkit.org/show_bug.cgi?id=145014

Reviewed by Martin Robinson.

.:

Error out immediately after checking for Ruby if the ruby executable is not found, or if it
is too old.

* CMakeLists.txt:

Source/JavaScriptCore:

Don't enforce the check for the Ruby executable here; it's now enforced in the top-level
CMakeLists.txt instead.

* CMakeLists.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184363 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed build fix after r184359; typo.
jer.noble@apple.com [Thu, 14 May 2015 23:52:29 +0000 (23:52 +0000)]
Unreviewed build fix after r184359; typo.

* html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::layoutSizeChanged):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184362 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoAdjust button CSS and positioning in preparation.
roger_fong@apple.com [Thu, 14 May 2015 23:40:29 +0000 (23:40 +0000)]
Adjust button CSS and positioning in preparation.
https://bugs.webkit.org/show_bug.cgi?id=144973.
<rdar://problem/20306227>

Reviewed by Dean Jackson.

The only visual change here is the swapping of the rewind and play button positions.
Also, position buttons based off of both left and right margins instead of just one of the two.
This allows the controls drop off to work without having to use a spacer element to take the place
of the timeline if the controls are too small.
* Modules/mediacontrols/mediaControlsApple.css:
(audio::-webkit-media-controls-rewind-button):
(audio::-webkit-media-controls-play-button):
(audio::-webkit-media-controls-panel .mute-box):
(audio::-webkit-media-controls-wireless-playback-picker-button):
(audio::-webkit-media-controls-toggle-closed-captions-button):
(audio::-webkit-media-controls-fullscreen-button):
(audio::-webkit-media-controls-fullscreen-button.exit):
(audio::-webkit-media-controls-time-remaining-display):
(audio:-webkit-full-screen::-webkit-media-controls-toggle-closed-captions-button):
(audio:-webkit-full-screen::-webkit-media-controls-wireless-playback-picker-button):
* Modules/mediacontrols/mediaControlsApple.js:
(Controller.prototype.configureInlineControls):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184361 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoMediaControls: controls are live even when invisible
dino@apple.com [Thu, 14 May 2015 23:35:42 +0000 (23:35 +0000)]
MediaControls: controls are live even when invisible
https://bugs.webkit.org/show_bug.cgi?id=145029
<rdar://problem/20865442>

Reviewed by Jer Noble.

When the controls are invisible they should ignore touch/mouse
events.

* Modules/mediacontrols/mediaControlsiOS.css: Add pointer-events: none where appropriate.
(video::-webkit-media-controls-panel-container):
(video::-webkit-media-controls-panel-background):
(video::-webkit-media-controls-panel):
(video::-webkit-media-controls-panel.paused):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184360 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[MediaControls] Refactor media controls & bring improvements made to iOS controls...
jer.noble@apple.com [Thu, 14 May 2015 23:27:35 +0000 (23:27 +0000)]
[MediaControls] Refactor media controls & bring improvements made to iOS controls to Mac.
https://bugs.webkit.org/show_bug.cgi?id=144973

Reviewed by Dean Jackson.

Pull improvements made to the iOS media controls back into the Mac controls by moving
code from mediaControlsiOS.js into MediaControlsApple.js.

The largest refactored feature is the ability to drop individual controls from the media
controls when the video is too small to contain them. To allow these controls to resize
dynamically, a new "resize" event is fired inside the media element's shadow DOM.

* Modules/mediacontrols/mediaControlsApple.css:
(audio::-webkit-media-controls-panel .dropped): Added; sets "display: none".
* Modules/mediacontrols/mediaControlsApple.js:
(Controller): Set defaults for new variables.
(Controller.prototype.updateControls): Update the controls width; moved from iOS.js.
(Controller.prototype.handleReadyStateChange): Update the controls; moved from iOS.js.
(Controller.prototype.handleTimeUpdate): Update the progress; moved from iOS.js.
(Controller.prototype.handleTimelineInput): Pause if scrubbing; moved from iOS.js.
(Controller.prototype.handleTimelineChange): Update the progress; moved from iOS.js.
(Controller.prototype.showControls): Update the controls width; moved from iOS.js.
(Controller.prototype.hideControls): Removed _potentiallyScrubbing check; not needed due to changes
    to controlsAlwaysVisible().
(Controller.prototype.scheduleUpdateLayoutForDisplayedWidth): Moved from iOS.js.
(Controller.prototype.isControlVisible): Added; checks whether control is parented & not hidden.
(Controller.prototype.updateLayoutForDisplayedWidth): Moved from iOS.js and refactored.
(Controller.prototype.controlsAlwaysVisible): Return true if scrubbing.
(Controller.prototype.updateHasAudio): Check currentPlaybackTargetIsWireless(); moved from iOS.js.
(Controller.prototype.get scrubbing): Simple getter for _scrubbing.
(Controller.prototype.set scrubbing): Check play state if scrubbing; start playback (if necessary)
    if not scrubbing.
(Controller.prototype.get pageScaleFactor): Moved from iOS.js.
(Controller.prototype.set pageScaleFactor): Ditto.
(Controller.prototype.handleRootResize): Schedule an update of the contrtols width.

Remove a bunch of newly unnecessary code from the iOS media controls:

* Modules/mediacontrols/mediaControlsiOS.js:
(ControllerIOS):
(ControllerIOS.prototype.createControls): Remove ivars moved into Apple.js.
(ControllerIOS.prototype.configureInlineControls): Remove spacer; made unnecessary.
(ControllerIOS.prototype.showControls): Deleted.
(ControllerIOS.prototype.updateTime): Deleted.
(ControllerIOS.prototype.handleTimelineTouchStart): Just call "scrubbing = true", handled in Apple.js.
(ControllerIOS.prototype.handleTimelineTouchEnd): Just call "scrubbing = false", handled in Apple.js.
(ControllerIOS.prototype.handleReadyStateChange): Deleted.
(ControllerIOS.prototype.setPlaying): Don't check _timelineIsHidden; not needed.
(ControllerIOS.prototype.get pageScaleFactor): Deleted.
(ControllerIOS.prototype.set pageScaleFactor): Deleted.
(ControllerIOS.prototype.scheduleUpdateLayoutForDisplayedWidth): Deleted.
(ControllerIOS.prototypeupdateLayoutForDisplayedWidth): Deleted.

Fire a "resize" event at the shadow DOM root when layout results in a size change.

* html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::layoutSizeChanged): Fire the "resize" event at the shadow DOM.
* html/HTMLMediaElement.h:
* rendering/RenderMedia.cpp:
(WebCore::RenderMedia::layout): Trigger layoutSizeChanged()
* rendering/RenderMedia.h:

Drive-by fixes:

* Modules/mediacontrols/mediaControlsApple.js:
(Controller.prototype.createControls): aria-label text is totally wrong; removed.
(Controller.prototype.updateWirelessPlaybackStatus): Use class-names to hide controls, not inline styles.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184359 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoAdd a layout mode that scales down the view to try to fit the document
timothy_horton@apple.com [Thu, 14 May 2015 22:46:15 +0000 (22:46 +0000)]
Add a layout mode that scales down the view to try to fit the document
https://bugs.webkit.org/show_bug.cgi?id=145022
<rdar://problem/19790341>

Reviewed by Dean Jackson.

* Shared/WebPageCreationParameters.cpp:
(WebKit::WebPageCreationParameters::encode):
(WebKit::WebPageCreationParameters::decode):
* Shared/WebPageCreationParameters.h:
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::creationParameters):
(WebKit::WebPageProxy::setShouldScaleViewToFitDocument):
* UIProcess/WebPageProxy.h:
* WebProcess/WebPage/DrawingArea.h:
(WebKit::DrawingArea::setShouldScaleViewToFitDocument):
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::WebPage):
(WebKit::WebPage::setShouldScaleViewToFitDocument):
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in:
Plumb shouldScaleViewToFitDocument through to the DrawingArea.

* UIProcess/mac/WKViewLayoutStrategy.mm:
(+[WKViewLayoutStrategy layoutStrategyWithPage:view:mode:]):
(-[WKViewDynamicSizeComputedFromMinimumDocumentSizeLayoutStrategy initWithPage:view:mode:]):
(-[WKViewDynamicSizeComputedFromMinimumDocumentSizeLayoutStrategy updateLayout]):
(-[WKViewDynamicSizeComputedFromMinimumDocumentSizeLayoutStrategy willChangeLayoutStrategy]):
* UIProcess/API/C/WKLayoutMode.h:
* UIProcess/API/Cocoa/_WKLayoutMode.h:
Add a new layout mode, which just turns on shouldScaleViewToFitDocument,
and otherwise behaves as normal.

* WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.h:
* WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.mm:
(WebKit::TiledCoreAnimationDrawingArea::setShouldScaleViewToFitDocument):
(WebKit::TiledCoreAnimationDrawingArea::scaleViewToFitDocumentIfNeeded):
(WebKit::TiledCoreAnimationDrawingArea::flushLayers):
On every flush where either the document size or view size has changed,
or layout is outstanding, do a layout with fixed layout off to determine
whether the document fits inside the view. If it doesn't, scale it down
to fit. This will require an extra layout for every resize while in the
scaled-down state, but there is potential for future optimization.

* MiniBrowser/mac/BrowserWindow.xib:
* MiniBrowser/mac/BrowserWindowController.h:
* MiniBrowser/mac/WK2BrowserWindowController.m:
(-[WK2BrowserWindowController toggleShrinkToFit:]):
(-[WK2BrowserWindowController toggleUseMinimumViewSize:]): Deleted.
Switch to _WKLayoutModeDynamicSizeComputedFromMinimumDocumentSize.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184358 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[CMake] Don't read the LOCATION property of targets
mcatanzaro@igalia.com [Thu, 14 May 2015 22:29:25 +0000 (22:29 +0000)]
[CMake] Don't read the LOCATION property of targets
https://bugs.webkit.org/show_bug.cgi?id=145018

Reviewed by Martin Robinson.

Use the TARGET_FILE_DIR generator expression to determine the location of the test injected
bundle, rather than assuming that the LOCATION property of TestWebKitAPIInjectedBundle will
be the same at configure-time as it is at generate-time.

* TestWebKitAPI/CMakeLists.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184357 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoLocal storage origins should include origins with transient local storage
andersca@apple.com [Thu, 14 May 2015 21:43:30 +0000 (21:43 +0000)]
Local storage origins should include origins with transient local storage
https://bugs.webkit.org/show_bug.cgi?id=145017
rdar://problem/10690447

Reviewed by Sam Weinig.

The transient local storage namespaces are used for third party data blocking and will stay
around until the UI process exits so we need to be able to include website data from transient storage
in the website data store APIs.

* UIProcess/Storage/StorageManager.cpp:
(WebKit::StorageManager::TransientLocalStorageNamespace::origins):
(WebKit::StorageManager::getLocalStorageOrigins):
* UIProcess/Storage/StorageManager.h:
* UIProcess/WebKeyValueStorageManager.cpp:
(WebKit::WebKeyValueStorageManager::getKeyValueStorageOrigins):
* UIProcess/WebsiteData/WebsiteDataStore.cpp:
(WebKit::WebsiteDataStore::fetchData):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184356 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoCrash in ReplaceSelectionCommand::removeRedundantStylesAndKeepStyleSpanInline
rniwa@webkit.org [Thu, 14 May 2015 21:39:50 +0000 (21:39 +0000)]
Crash in ReplaceSelectionCommand::removeRedundantStylesAndKeepStyleSpanInline
https://bugs.webkit.org/show_bug.cgi?id=119068

Reviewed by Enrica Casucci.

Source/WebCore:

The bug was caused by makeInsertedContentRoundTrippableWithHTMLTreeBuilder not updating
nodes kept tracked by insertedNodes and moveNodeOutOfAncestor stumbling upon it.

Fixed the bug by updating insertedNodes in makeInsertedContentRoundTrippableWithHTMLTreeBuilder.

Test: editing/inserting/insert-table-in-paragraph-crash.html

* editing/ReplaceSelectionCommand.cpp:
(WebCore::ReplaceSelectionCommand::makeInsertedContentRoundTrippableWithHTMLTreeBuilder):
(WebCore::ReplaceSelectionCommand::moveNodeOutOfAncestor):
* editing/ReplaceSelectionCommand.h:

LayoutTests:

Added a test based on https://chromium.googlesource.com/chromium/blink/+/3500267482e60550ce84fadd6c0db883937ce744

* editing/inserting/insert-table-in-paragraph-crash-expected.txt: Added.
* editing/inserting/insert-table-in-paragraph-crash.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184355 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoEnforce options coherency
basile_clement@apple.com [Thu, 14 May 2015 21:32:05 +0000 (21:32 +0000)]
Enforce options coherency
https://bugs.webkit.org/show_bug.cgi?id=144921

Reviewed by Mark Lam.

JavaScriptCore should be failing early when the options are set in such
a way that we don't have a meaningful way to execute JavaScript, rather
than failing for obscure reasons at some point during execution.

This patch adds a new function that checks whether the options are set
in a coherent way, and makes JSC::Options::initialize() crash when the
environment enforces incoherent options.
Client applications able to add or change additional options are
responsible to check for coherency again before starting to actually
execute JavaScript, if any additional options have been set. This is
implemented for the jsc executable in this patch.

* jsc.cpp:
(CommandLine::parseArguments):
* runtime/Options.cpp:
(JSC::Options::initialize):
(JSC::Options::ensureOptionsAreCoherent): Added.
* runtime/Options.h:
(JSC::Options::ensureOptionsAreCoherent): Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184354 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[Mac] Expose more font weights for -apple-system
mmaxfield@apple.com [Thu, 14 May 2015 21:28:54 +0000 (21:28 +0000)]
[Mac] Expose more font weights for -apple-system
https://bugs.webkit.org/show_bug.cgi?id=144707

Reviewed by Simon Fraser.

Source/WebCore:

Previously, when we parsed a CSS declaration of the form font: keyword; where keyword
is one of caption, icon, menu, message-box, small-caption, -webkit-mini-control, -webkit-small-control,
or -webkit-control (which html.css does for form controls), we would ask the system what the appropriate
system font is, get that font's family name, and synthesize a font-family CSS property for the element.
Then, later when we actually go to look up the font, we would look up the font by family name using this
information. However, this round-tripping of a font through a family name is actually lossy, and is not
guaranteed to preserve system-font-ness (which we use for various things including metrics calculations).

This patch modifies this logic to specify a token family name instead, which the font lookup code special
cases (and reacts by making the appropriate system-font lookup call). This approach is currently how iOS
handles these system fonts; this patch simply brings this approach to OS X.

There is also an added progression here. We used to simply call [NSFont fontWithName:size:] on the system
font family name (which the parser found for us) which entirely disregards weight. This means that we
used to be getting synthesized bold in form controls which ask for a heavy weight. Migrating to this
system-font aware call means that we get the real bold font instead of synthesized bold.

Once this system-font-ness is guaranteed to be preserved between parsing time and font lookup time, we
can safely migrate to using [NSFont systemFontOfSize:weight] instead of [NSFont systemFontOfSize:] on
platforms which support it.

Tests: fast/text/systemFont.html
       fast/css/css2-system-fonts.html
       fast/forms/select/optgroup-rendering.html
       fast/forms/validation-message-appearance.html

* css/CSSParser.cpp:
(WebCore::CSSParser::parseSystemFont): Add a comment regarding why we are bothering with expanding out
the font property in the first place.
* platform/graphics/cocoa/FontCascadeCocoa.mm:
(WebCore::FontCascade::primaryFontIsSystemFont): Update to use new system font tokens.
* platform/graphics/mac/FontCacheMac.mm:
(WebCore::toNSFontWeight): New static method to map font weights to NSFontWeight constants available on
Yosemite and later.
(WebCore::fontWithFamilySpecialCase): Pull all these special-case font token name handling into a
separate function, which returns an Optional.
(WebCore::fontWithFamily):
* platform/mac/ThemeMac.mm:
(WebCore::ThemeMac::controlFont): Use the font token name instead of the generated system font family
name.
* platform/spi/mac/NSFontSPI.h: Add [NSFont systemFontWithSize:weight:] and the proper NSFontWeight
constants.
* rendering/RenderThemeMac.mm:
(WebCore::RenderThemeMac::updateCachedSystemFontDescription): Use the font token names instead of the
generated system font family name.
(WebCore::RenderThemeMac::setFontFromControlSize): Ditto.

LayoutTests:

* platform/mac/fast/text/systemFont-expected.txt: Update expectations.
* platform/mac/fast/text/systemFont.html: Update test to include font weights for -apple-system.
* platform/mac/fast/css/css2-system-fonts-expected.txt: Updated to not hardcode the system font family name.
* platform/mac-mavericks/fast/css/css2-system-fonts-expected.txt: Ditto.
* platform/mac/fast/forms/select/optgroup-rendering-expected.txt: Updated to not use synthetic bold.
* platform/mac/fast/forms/validation-message-appearance-expected.txt: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184353 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoREGRESSION (r184337): [EFL] unresolved reference errors in ARM builds
utatane.tea@gmail.com [Thu, 14 May 2015 21:24:35 +0000 (21:24 +0000)]
REGRESSION (r184337): [EFL] unresolved reference errors in ARM builds
https://bugs.webkit.org/show_bug.cgi?id=145019

Reviewed by Ryosuke Niwa.

Attempt to fix compile errors in EFL ARM buildbots.
By executing `nm`, found JSTemplateRegistryKey.cpp.o and TemplateRegistry.cpp.o have
unresolved reference to Structure::get. That is inlined function in StructureInlines.h.

* runtime/JSTemplateRegistryKey.cpp:
* runtime/TemplateRegistry.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184352 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoAdd internals setting to disable wireless playback availability for layout tests
roger_fong@apple.com [Thu, 14 May 2015 21:19:09 +0000 (21:19 +0000)]
Add internals setting to disable wireless playback availability for layout tests
https://bugs.webkit.org/show_bug.cgi?id=145012.
<rdar://problem/20946504>

Reviewed by Eric Carlson.

* testing/InternalSettings.cpp:
(WebCore::InternalSettings::resetToConsistentState):
(WebCore::InternalSettings::setWirelessPlaybackDisabled):
* testing/InternalSettings.idl:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184351 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoSmall refactoring before implementation of the ES6 arrow function.
commit-queue@webkit.org [Thu, 14 May 2015 20:38:43 +0000 (20:38 +0000)]
Small refactoring before implementation of the ES6 arrow function.
https://bugs.webkit.org/show_bug.cgi?id=144954

Patch by Alexandr Skachkov <gskachkov@gmail.com> on 2015-05-14
Reviewed by Ryosuke Niwa.

* parser/Parser.h:
* parser/Parser.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184349 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoREGRESSION (r184337): ASSERT failed in debug builds for tagged templates
utatane.tea@gmail.com [Thu, 14 May 2015 19:58:00 +0000 (19:58 +0000)]
REGRESSION (r184337): ASSERT failed in debug builds for tagged templates
https://bugs.webkit.org/show_bug.cgi?id=145013

Reviewed by Filip Pizlo.

Fix the regression introduced by r184337.

1. JSTemporaryRegistryKey::s_info should inherit the Base::s_info,
   JSDestructibleObject::s_info.

2. The first register argument of BytecodeGenerator::emitNode
   should be a referenced register if it is a temporary register.

* bytecompiler/NodesCodegen.cpp:
(JSC::TaggedTemplateNode::emitBytecode):
* runtime/JSTemplateRegistryKey.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184347 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoString.prototype.split() should create efficient substrings.
akling@apple.com [Thu, 14 May 2015 19:07:30 +0000 (19:07 +0000)]
String.prototype.split() should create efficient substrings.
<https://webkit.org/b/144985>
<rdar://problem/20949344>

Reviewed by Geoffrey Garen.

Teach split() how to make substring JSStrings instead of relying on StringImpl's
substring sharing mechanism. The optimization works by deferring the construction
of a StringImpl until the substring's value is actually needed.

This knocks ~2MB off of theverge.com by avoiding the extra StringImpl allocations.
Out of ~70000 substrings created by split(), only ~2000 of them get reified.

* runtime/StringPrototype.cpp:
(JSC::jsSubstring):
(JSC::splitStringByOneCharacterImpl):
(JSC::stringProtoFuncSplit):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184346 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoChange range of possible forces for mouseforcechanged DOM event
bdakin@apple.com [Thu, 14 May 2015 18:17:39 +0000 (18:17 +0000)]
Change range of possible forces for mouseforcechanged DOM event
https://bugs.webkit.org/show_bug.cgi?id=144987
-and corresponding-
rdar://problem/20472802

Reviewed by Tim Horton.

Change to a 0-3 range.
Source/WebCore:

* platform/PlatformMouseEvent.h:
* platform/mac/PlatformEventFactoryMac.mm:
(WebCore::PlatformMouseEventBuilder::PlatformMouseEventBuilder):

Source/WebKit2:

* Shared/mac/WebEventFactory.mm:
(WebKit::WebEventFactory::createWebMouseEvent):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184345 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoChange the status of ES6 tagged templates to Done in features.json
utatane.tea@gmail.com [Thu, 14 May 2015 18:11:55 +0000 (18:11 +0000)]
Change the status of ES6 tagged templates to Done in features.json
https://bugs.webkit.org/show_bug.cgi?id=145003

Reviewed by Benjamin Poulain.

Now it's implemented in r184337.

* features.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184344 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoAdd String literal overloads to equalIgnoringASCIICase()
mmaxfield@apple.com [Thu, 14 May 2015 17:55:38 +0000 (17:55 +0000)]
Add String literal overloads to equalIgnoringASCIICase()
https://bugs.webkit.org/show_bug.cgi?id=145008

Patch by Myles C. Maxfield <mmaxfield@apple.com> on 2015-05-14
Reviewed by Benjamin Poulain.

Source/WTF:

Create an overload for equalIgnoringASCIICase for string literals.

* wtf/text/StringImpl.h:
(WTF::equalIgnoringASCIICase): Use a non-templated helper function.
* wtf/text/StringImpl.cpp:
(WTF::equalIgnoringASCIICase): Implement it.
* wtf/text/StringView.h:
(WTF::equalIgnoringASCIICase): Use a non-templated helper function.
* wtf/text/StringView.cpp:
(WTF::equalIgnoringASCIICase): Implement it.
* wtf/text/WTFString.h:
(WTF::equalIgnoringASCIICase): Delegate to StringImpl's implementation.

Tools:

Test changes to WTF.

* TestWebKitAPI/Tests/WTF/StringImpl.cpp:
(WTF.StringImplEqualIgnoringASCIICaseBasic): Test const char*.
(WTF.StringImplEqualIgnoringASCIICaseWithLatin1Characters): Ditto.
* TestWebKitAPI/Tests/WTF/StringView.cpp:
(WTF.StringViewEqualIgnoringASCIICaseBasic): Ditto.
(WTF.StringViewEqualIgnoringASCIICaseWithLatin1Characters): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184341 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoIntroduce SymbolType into SpeculativeTypes
utatane.tea@gmail.com [Thu, 14 May 2015 17:36:12 +0000 (17:36 +0000)]
Introduce SymbolType into SpeculativeTypes
https://bugs.webkit.org/show_bug.cgi?id=142651

Reviewed by Filip Pizlo.

Introduce SpecSymbol type into speculative types.
Previously symbol type is categorized into SpecCellOther.
But SpecCellOther is not intended to be used for such cells.

This patch just introduces SpecSymbol.
It represents the type of target value is definitely the symbol type.
It is the part of SpecCell.

In this patch, we do not introduce SymbolUse tracking.
It will be added in the separate patch.

* bytecode/SpeculatedType.cpp:
(JSC::dumpSpeculation):
(JSC::speculationFromStructure):
* bytecode/SpeculatedType.h:
(JSC::isSymbolSpeculation):
* dfg/DFGAbstractInterpreterInlines.h:
(JSC::DFG::AbstractInterpreter<AbstractStateType>::executeEffects):
* dfg/DFGAbstractValue.cpp:
(JSC::DFG::AbstractValue::setType):
* dfg/DFGConstantFoldingPhase.cpp:
(JSC::DFG::ConstantFoldingPhase::foldConstants):
* tests/stress/typeof-symbol.js: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184340 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoFix typo in RenderBox::instrinsicScrollbarLogicalWidth()
rego@igalia.com [Thu, 14 May 2015 16:37:27 +0000 (16:37 +0000)]
Fix typo in RenderBox::instrinsicScrollbarLogicalWidth()
https://bugs.webkit.org/show_bug.cgi?id=144999

Reviewed by Sergio Villar Senin.

Rename RenderBox::instrinsicScrollbarLogicalWidth() to
RenderBox::intrinsicScrollbarLogicalWidth().

No new tests, no behavior changes.

* rendering/RenderBlock.cpp:
(WebCore::RenderBlock::computeIntrinsicLogicalWidths):
* rendering/RenderBlockFlow.cpp:
(WebCore::RenderBlockFlow::computeIntrinsicLogicalWidths):
* rendering/RenderBox.cpp:
(WebCore::RenderBox::intrinsicScrollbarLogicalWidth):
(WebCore::RenderBox::instrinsicScrollbarLogicalWidth): Deleted.
* rendering/RenderBox.h:
* rendering/RenderDeprecatedFlexibleBox.cpp:
(WebCore::RenderDeprecatedFlexibleBox::computeIntrinsicLogicalWidths):
* rendering/RenderFlexibleBox.cpp:
(WebCore::RenderFlexibleBox::computeIntrinsicLogicalWidths):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184339 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed, skip js/regress-141098.html. The fix will be tracked in https://bugs...
fpizlo@apple.com [Thu, 14 May 2015 16:31:23 +0000 (16:31 +0000)]
Unreviewed, skip js/regress-141098.html. The fix will be tracked in https://bugs.webkit.org/show_bug.cgi?id=145007

* TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184338 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[ES6] Implement tagged templates
utatane.tea@gmail.com [Thu, 14 May 2015 16:07:54 +0000 (16:07 +0000)]
[ES6] Implement tagged templates
https://bugs.webkit.org/show_bug.cgi?id=143183

Reviewed by Oliver Hunt.

This patch implements ES6 tagged templates.
In tagged templates, the function takes the template object.

The template object contains the raw and cooked template strings,
so when parsing the tagged templates, we need to tokenize the raw and cooked strings.
While tagged templates require the both strings, the template literal only requires
the cooked strings. So when tokenizing under the template literal context,
we only builds the cooked strings.

As per ES6 spec, the template objects for the same raw strings are shared in the same realm.
The template objects is cached. And every time we evaluate the same tagged templates,
the same (cached) template objects are used.
Since the spec freezes this template objects completely,
we cannot attach some properties to it.
So we can say that it behaves as if the template objects are the primitive values (like JSString).
Since we cannot attach properties, the only way to test the identity of the template object is comparing. (===)
As the result, when there is no reference to the template object, we can garbage collect it
because the user has no way to test that the newly created template object does not equal
to the already collected template object.

So, to implement tagged templates, we implement the following components.

1. JSTemplateRegistryKey
It holds the template registry key and it does not exposed to users.
TemplateRegistryKey holds the vector of raw and cooked strings with the pre-computed hash value.
When obtaining the template object for the (statically, a.k.a. at the parsing time) given raw string vectors,
we use this JSTemplateRegistryKey as a key to the map and look up the template object from
TemplateRegistry.
JSTemplateRegistryKey is created at the bytecode compiling time and
stored in the CodeBlock as like as JSString content values.

2. TemplateRegistry
This manages the cached template objects.
It holds the weak map (JSTemplateRegistryKey -> the template object).
The template object is weakly referenced.
So if there is no reference to the template object,
the template object is automatically GC-ed.
When looking up the template object, it searches the cached template object.
If it is found, it is returned to the users.
If there is no cached template objects, it creates the new template object and
stores it with the given template registry key.

* CMakeLists.txt:
* JavaScriptCore.vcxproj/JavaScriptCore.vcxproj:
* JavaScriptCore.vcxproj/JavaScriptCore.vcxproj.filters:
* JavaScriptCore.xcodeproj/project.pbxproj:
* bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::addTemplateRegistryKeyConstant):
(JSC::BytecodeGenerator::emitGetTemplateObject):
* bytecompiler/BytecodeGenerator.h:
* bytecompiler/NodesCodegen.cpp:
(JSC::TaggedTemplateNode::emitBytecode):
(JSC::TemplateLiteralNode::emitBytecode): Deleted.
* parser/ASTBuilder.h:
(JSC::ASTBuilder::createTaggedTemplate):
(JSC::ASTBuilder::createTemplateLiteral): Deleted.
* parser/Lexer.cpp:
(JSC::Lexer<T>::setCode):
(JSC::Lexer<T>::parseTemplateLiteral):
(JSC::Lexer<T>::lex):
(JSC::Lexer<T>::scanTrailingTemplateString):
(JSC::Lexer<T>::clear):
* parser/Lexer.h:
(JSC::Lexer<T>::makeEmptyIdentifier):
* parser/NodeConstructors.h:
(JSC::TaggedTemplateNode::TaggedTemplateNode):
(JSC::TemplateLiteralNode::TemplateLiteralNode): Deleted.
* parser/Nodes.h:
(JSC::TemplateLiteralNode::templateStrings):
(JSC::TemplateLiteralNode::templateExpressions):
(JSC::TaggedTemplateNode::templateLiteral):
* parser/Parser.cpp:
(JSC::Parser<LexerType>::parseTemplateString):
(JSC::Parser<LexerType>::parseTemplateLiteral):
(JSC::Parser<LexerType>::parsePrimaryExpression):
(JSC::Parser<LexerType>::parseMemberExpression):
* parser/Parser.h:
* parser/ParserArena.h:
(JSC::IdentifierArena::makeEmptyIdentifier):
* parser/SyntaxChecker.h:
(JSC::SyntaxChecker::createTaggedTemplate):
(JSC::SyntaxChecker::createTemplateLiteral): Deleted.
* runtime/CommonIdentifiers.h:
* runtime/JSGlobalObject.cpp:
(JSC::getTemplateObject):
(JSC::JSGlobalObject::JSGlobalObject):
(JSC::JSGlobalObject::init):
* runtime/JSGlobalObject.h:
(JSC::JSGlobalObject::templateRegistry):
* runtime/JSTemplateRegistryKey.cpp: Added.
(JSC::JSTemplateRegistryKey::JSTemplateRegistryKey):
(JSC::JSTemplateRegistryKey::create):
(JSC::JSTemplateRegistryKey::destroy):
* runtime/JSTemplateRegistryKey.h: Added.
* runtime/ObjectConstructor.cpp:
(JSC::objectConstructorFreeze):
* runtime/ObjectConstructor.h:
* runtime/TemplateRegistry.cpp: Added.
(JSC::TemplateRegistry::TemplateRegistry):
(JSC::TemplateRegistry::getTemplateObject):
* runtime/TemplateRegistry.h: Added.
* runtime/TemplateRegistryKey.h: Added.
(JSC::TemplateRegistryKey::isDeletedValue):
(JSC::TemplateRegistryKey::isEmptyValue):
(JSC::TemplateRegistryKey::hash):
(JSC::TemplateRegistryKey::rawStrings):
(JSC::TemplateRegistryKey::cookedStrings):
(JSC::TemplateRegistryKey::operator==):
(JSC::TemplateRegistryKey::operator!=):
(JSC::TemplateRegistryKey::Hasher::hash):
(JSC::TemplateRegistryKey::Hasher::equal):
(JSC::TemplateRegistryKey::TemplateRegistryKey):
* runtime/VM.cpp:
(JSC::VM::VM):
* runtime/VM.h:
* tests/stress/tagged-templates-identity.js: Added.
(shouldBe):
* tests/stress/tagged-templates-raw-strings.js: Added.
(shouldBe):
(tag):
(testEval):
* tests/stress/tagged-templates-syntax.js: Added.
(tag):
(testSyntax):
(testSyntaxError):
* tests/stress/tagged-templates-template-object.js: Added.
(shouldBe):
(tag):
* tests/stress/tagged-templates-this.js: Added.
(shouldBe):
(tag):
* tests/stress/tagged-templates.js: Added.
(shouldBe):
(raw):
(cooked):
(Counter):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184337 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWeb Inspector: Current time marker is always at zero in Rendering Frames ruler
mattbaker@apple.com [Thu, 14 May 2015 15:33:12 +0000 (15:33 +0000)]
Web Inspector: Current time marker is always at zero in Rendering Frames ruler
https://bugs.webkit.org/show_bug.cgi?id=144518

Reviewed by Timothy Hatcher.

The current and end time values for the rendering frame timeline overview should always be equal to the frame
number of the last record in the rendering frames timeline.

* UserInterface/Views/TimelineOverview.js:
(WebInspector.TimelineOverview):
(WebInspector.TimelineOverview.prototype.updateLayout):
* UserInterface/Views/TimelineRecordingContentView.js:
(WebInspector.TimelineRecordingContentView.prototype._updateTimes):
(WebInspector.TimelineRecordingContentView.prototype._recordingTimesUpdated):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184336 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[GTK] Enable plugin-related CMake options and variables for the X11 target only
zandobersek@gmail.com [Thu, 14 May 2015 10:52:20 +0000 (10:52 +0000)]
[GTK] Enable plugin-related CMake options and variables for the X11 target only
https://bugs.webkit.org/show_bug.cgi?id=144995

Reviewed by Carlos Garcia Campos.

* Source/cmake/OptionsGTK.cmake: Plugins are only supported for
the X11 windowing target at the moment, so the following options
and variables should be enabled or disabled accordingly:
- ENABLE_PLUGIN_PROCESS_GTK2
- ENABLE_NETSCAPE_PLUGIN_API
- ENABLE_PLUGIN_PROCESS

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@184335 268f45cc-cd09-0410-ab3c-d52691b4dbfc