WebKit-https.git
2 years ago[GStreamer] Use smart pointers and modernize code in WebKitWebAudioSourceGStreamer
carlosgc@webkit.org [Wed, 11 Jan 2017 11:27:38 +0000 (11:27 +0000)]
[GStreamer] Use smart pointers and modernize code in WebKitWebAudioSourceGStreamer
https://bugs.webkit.org/show_bug.cgi?id=166886

Reviewed by Xabier Rodriguez-Calvar.

This patch doesn't change the behavior, so it's covered by existing Web Audio tests. It replaces pointers with
smart pointers, uses WTF::Vector instead of GSList and simplifies the code to map/unmap GstBuffers.

* platform/audio/gstreamer/WebKitWebAudioSourceGStreamer.cpp:
(webKitWebAudioSrcConstructed):
(webKitWebAudioSrcFinalize):
(webKitWebAudioSrcLoop):
(webKitWebAudioSrcChangeState):
* platform/graphics/gstreamer/GRefPtrGStreamer.cpp:
(WTF::derefGPtr<GstBufferList>):
(WTF::adoptGRef):
(WTF::refGPtr<GstBufferPool>):
(WTF::derefGPtr<GstBufferPool>):
* platform/graphics/gstreamer/GRefPtrGStreamer.h:
* platform/graphics/gstreamer/GStreamerUtilities.cpp:
(WebCore::mapGstBuffer):
* platform/graphics/gstreamer/GStreamerUtilities.h:
* platform/graphics/gstreamer/WebKitWebSourceGStreamer.cpp:
(StreamingClient::createReadBuffer):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210584 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoUnreviewed, rolling out r182947.
commit-queue@webkit.org [Wed, 11 Jan 2017 09:08:11 +0000 (09:08 +0000)]
Unreviewed, rolling out r182947.
https://bugs.webkit.org/show_bug.cgi?id=166922

Caused huge memory leak see https://webkit.org/b/166884
(Requested by KaL on #webkit).

Reverted changeset:

"[GStreamer] Silent WebAudio buffers support"
https://bugs.webkit.org/show_bug.cgi?id=143869
http://trac.webkit.org/changeset/182947

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210583 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoUnreviewed, rolling out r210579.
carlosgc@webkit.org [Wed, 11 Jan 2017 08:55:22 +0000 (08:55 +0000)]
Unreviewed, rolling out r210579.

That was not the right fix

Reverted changeset:

"Unreviewed GTK buildfix after r210571."
http://trac.webkit.org/changeset/210579

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210582 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[EFL] Update TestExpectation file for Ubuntu 16.04.
gyuyoung.kim@webkit.org [Wed, 11 Jan 2017 08:53:05 +0000 (08:53 +0000)]
[EFL] Update TestExpectation file for Ubuntu 16.04.

Unreviewed EFL gardening.

After bumping Ubuntu version to 16.04, there are many tests are failed.
This gardening is to update all failure tests on Ubuntu 16.04.

* platform/efl/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210581 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoUnreviewed. Fix GTK+ build after r210571.
carlosgc@webkit.org [Wed, 11 Jan 2017 08:52:56 +0000 (08:52 +0000)]
Unreviewed. Fix GTK+ build after r210571.

Add getFileDeviceId implementation to FileSystemGlib.cpp.

* platform/glib/FileSystemGlib.cpp:
(WebCore::getFileDeviceId):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210580 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoUnreviewed GTK buildfix after r210571.
ossy@webkit.org [Wed, 11 Jan 2017 08:47:42 +0000 (08:47 +0000)]
Unreviewed GTK buildfix after r210571.

* PlatformGTK.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210579 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoAnother build fix. Always use UTC when expressing commit times in UNIX-epoch timestamps.
rniwa@webkit.org [Wed, 11 Jan 2017 07:53:17 +0000 (07:53 +0000)]
Another build fix. Always use UTC when expressing commit times in UNIX-epoch timestamps.

* public/api/measurement-set.php:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210578 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoFix a typo in the previous commit.
rniwa@webkit.org [Wed, 11 Jan 2017 04:53:53 +0000 (04:53 +0000)]
Fix a typo in the previous commit.

* public/api/measurement-set.php:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210577 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoBuild fixes for older versions of Postgres.
rniwa@webkit.org [Wed, 11 Jan 2017 04:51:12 +0000 (04:51 +0000)]
Build fixes for older versions of Postgres.

Also redirect / and /# to /v3/ as intended in r200820.

* public/api/measurement-set.php:
* public/api/runs.php:
* public/index.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210576 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[GTK] install-dependencies script misses gdb
clopez@igalia.com [Wed, 11 Jan 2017 04:07:54 +0000 (04:07 +0000)]
[GTK] install-dependencies script misses gdb
https://bugs.webkit.org/show_bug.cgi?id=166917

Reviewed by Michael Catanzaro.

gdb is needed for generating the crash logs.

* gtk/install-dependencies:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210575 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[GTK] No way for applications to set notification permissions without waiting for...
mcatanzaro@igalia.com [Wed, 11 Jan 2017 03:56:12 +0000 (03:56 +0000)]
[GTK] No way for applications to set notification permissions without waiting for permission request
https://bugs.webkit.org/show_bug.cgi?id=163366

Reviewed by Carlos Garcia Campos.

Source/WebKit2:

Websites can check the Notification.permission DOM property to check if they have permission
to show notifications *without triggering a permission request*. But the WebKitGTK+ API has
no way to affirmatively indicate notifications are allowed without a permission request. The
only way is via the permission request API, which is too late. It's a problem for Epiphany.
For example, open the Riot Matrix client in a web app; you will immediately see an info bar
informing the user that Riot does not have permission to send desktop notifications, even
though Epiphany automatically grants notification permission in web app mode when requested.
This problem is not exclusive to web apps; there is simply no way for notification
permission to be set correctly prior to a permission request for it. Fix this by introducing
a webkit_web_context_initialize_notification_permissions() API, and add a signal
WebKitWebContext::initialize-notification-permissions to indicate when it should be called.

* UIProcess/API/gtk/WebKitNotificationProvider.cpp:
(notificationPermissionsCallback):
(WebKitNotificationProvider::create):
(WebKitNotificationProvider::WebKitNotificationProvider):
(WebKitNotificationProvider::notificationPermissions):
(WebKitNotificationProvider::setNotificationPermissions):
* UIProcess/API/gtk/WebKitNotificationProvider.h:
* UIProcess/API/gtk/WebKitSecurityOrigin.cpp:
(webkitSecurityOriginGetSecurityOrigin):
* UIProcess/API/gtk/WebKitSecurityOriginPrivate.h:
* UIProcess/API/gtk/WebKitWebContext.cpp:
(webkitWebContextConstructed):
(webkit_web_context_class_init):
(addOriginToMap):
(webkit_web_context_initialize_notification_permissions):
(webkitWebContextInitializeNotificationPermissions):
* UIProcess/API/gtk/WebKitWebContext.h:
* UIProcess/API/gtk/WebKitWebContextPrivate.h:
* UIProcess/API/gtk/docs/webkit2gtk-4.0-sections.txt:

Tools:

* TestWebKitAPI/Tests/WebKit2Gtk/TestWebKitWebView.cpp:
(testWebViewNotification):
(setInitialNotificationPermissionsAllowedCallback):
(setInitialNotificationPermissionsDisallowedCallback):
(testWebViewNotificationInitialPermissionAllowed):
(testWebViewNotificationInitialPermissionDisallowed):
(beforeAll):
* TestWebKitAPI/gtk/WebKit2Gtk/TestMain.h:
* TestWebKitAPI/gtk/WebKit2Gtk/WebViewTest.cpp:
(WebViewTest::WebViewTest):
(WebViewTest::initializeWebView):
* TestWebKitAPI/gtk/WebKit2Gtk/WebViewTest.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210574 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoImplement JSSourceCode to propagate SourceCode in module pipeline
utatane.tea@gmail.com [Wed, 11 Jan 2017 02:26:47 +0000 (02:26 +0000)]
Implement JSSourceCode to propagate SourceCode in module pipeline
https://bugs.webkit.org/show_bug.cgi?id=166861

Reviewed by Saam Barati.

Source/JavaScriptCore:

Instead of propagating source code string, we propagate JSSourceCode
cell in the module pipeline. This allows us to attach a metadata
to the propagated source code string. In particular, it propagates
SourceOrigin through the module pipeline.

And it also fixes JSC shell to use Module source type for module source code.

* CMakeLists.txt:
* JavaScriptCore.xcodeproj/project.pbxproj:
* builtins/ModuleLoaderPrototype.js:
(fulfillFetch):
(requestFetch):
* jsc.cpp:
(GlobalObject::moduleLoaderFetch):
(runWithScripts):
* llint/LLIntData.cpp:
(JSC::LLInt::Data::performAssertions):
* llint/LowLevelInterpreter.asm:
* runtime/Completion.cpp:
(JSC::loadAndEvaluateModule):
(JSC::loadModule):
* runtime/JSModuleLoader.cpp:
(JSC::JSModuleLoader::provide):
* runtime/JSModuleLoader.h:
* runtime/JSSourceCode.cpp: Added.
(JSC::JSSourceCode::destroy):
* runtime/JSSourceCode.h: Added.
(JSC::JSSourceCode::createStructure):
(JSC::JSSourceCode::create):
(JSC::JSSourceCode::sourceCode):
(JSC::JSSourceCode::JSSourceCode):
* runtime/JSType.h:
* runtime/ModuleLoaderPrototype.cpp:
(JSC::moduleLoaderPrototypeParseModule):
* runtime/VM.cpp:
(JSC::VM::VM):
* runtime/VM.h:

Source/WebCore:

Now SourceOrigin is correctly propagated through the module pipeline.
So the error reported by http/tests/security/contentSecurityPolicy/module-eval-blocked-expected.html
is updated.

* ForwardingHeaders/runtime/JSSourceCode.h: Added.
* bindings/js/JSDOMPromise.h:
(WebCore::DeferredPromise::resolveWithCallback):
* bindings/js/ScriptModuleLoader.cpp:
(WebCore::ScriptModuleLoader::notifyFinished):

LayoutTests:

* http/tests/security/contentSecurityPolicy/module-eval-blocked-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210573 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoMobileMiniBrowser tests don't work by default
timothy_horton@apple.com [Wed, 11 Jan 2017 01:54:25 +0000 (01:54 +0000)]
MobileMiniBrowser tests don't work by default
https://bugs.webkit.org/show_bug.cgi?id=166910

Reviewed by Alex Christensen.

* MobileMiniBrowser/MobileMiniBrowser.xcodeproj/xcshareddata/xcschemes/MobileMiniBrowserUITests.xcscheme: Added.
Add a shared scheme; the Makefile depends on the MobileMiniBrowserUITests scheme existing, at the very least.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210572 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoFile scheme should not allow access of a resource on a different volume.
bfulgham@apple.com [Wed, 11 Jan 2017 01:51:25 +0000 (01:51 +0000)]
File scheme should not allow access of a resource on a different volume.
https://bugs.webkit.org/show_bug.cgi?id=158552
<rdar://problem/15307582>

Reviewed by Alex Christensen.

Source/WebCore:

Revise SecurityOrigin to prevent files from one storage device (volume) from accessing content
on a different storage device (volume) unless universal access is enabled.

Pass the current file device as part of the NSURLRequest so that CFNetwork can reject loads
where the device changes in the midst of a load.

Also properly reflect that SecurityOrigin is never null by passing as a reference,
rather than as a pointer.

Tests: Tools/TestWebKitAPI/Tests/mac/CrossPartitionFileSchemeAccess.mm

* page/SecurityOrigin.cpp:
(WebCore::SecurityOrigin::canAccess): Pass argument as reference.
(WebCore::SecurityOrigin::passesFileCheck): Add check that file URLs refer to files in
the same storage volume.
(WebCore::SecurityOrigin::canDisplay): Add check that files share the same volume.
(WebCore::SecurityOrigin::isSameSchemeHostPort): Pass argument as reference.
* page/SecurityOrigin.h:
* platform/FileSystem.cpp:
(WebCore::filesHaveSameVolume): Added.
* platform/FileSystem.h:
* platform/network/cocoa/ResourceRequestCocoa.mm:
(WebCore::ResourceRequest::doUpdatePlatformRequest): If loading a file URL, tell CFNetwork
the storage device at the time of the start of the load so we can trigger a failure if this
changes during the load operation.
* platform/posix/FileSystemPOSIX.cpp:
(WebCore::getFileDeviceId): Added.
* platform/win/FileSystemWin.cpp:
(WebCore::getFileDeviceId): Added.

Tools:

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Add new files.
* TestWebKitAPI/Tests/mac/CrossPartitionFileSchemeAccess.html: Added.
* TestWebKitAPI/Tests/mac/CrossPartitionFileSchemeAccess.mm: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210571 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoUnreviewed, rolling out r210052.
commit-queue@webkit.org [Wed, 11 Jan 2017 01:35:32 +0000 (01:35 +0000)]
Unreviewed, rolling out r210052.
https://bugs.webkit.org/show_bug.cgi?id=166915

"breaks web compatability" (Requested by keith_miller on
#webkit).

Reverted changeset:

"Add support for global"
https://bugs.webkit.org/show_bug.cgi?id=165171
http://trac.webkit.org/changeset/210052

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210570 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoReverting run-jsc-stress-tests' $buildType to use string values.
mark.lam@apple.com [Wed, 11 Jan 2017 01:28:00 +0000 (01:28 +0000)]
Reverting run-jsc-stress-tests' $buildType to use string values.
https://bugs.webkit.org/show_bug.cgi?id=166912

Rubber stamped by Filip Pizlo.

This is so that we can test for it in the tests e.g.
//@ if $buildType == "debug" then runDefault("--maxSingleAllocationSize=1048576") else skip end

* Scripts/run-jsc-stress-tests:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210569 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[WebIDL] Remove most of the custom bindings for the WebGL code
commit-queue@webkit.org [Wed, 11 Jan 2017 01:02:34 +0000 (01:02 +0000)]
[WebIDL] Remove most of the custom bindings for the WebGL code
https://bugs.webkit.org/show_bug.cgi?id=166834

Patch by Sam Weinig <sam@webkit.org> on 2017-01-10
Reviewed by Alex Christensen.

Source/JavaScriptCore:

* runtime/ArrayPrototype.h:
* runtime/ObjectPrototype.h:
Export the ClassInfo so it can be used from WebCore.

Source/WebCore:

Removes all but one custom getter (getExtension) from the WebGL code.

* Modules/indexeddb/IDBKeyRange.idl:
Replace use of non-standard type, with new OverrideIDLType extended attribute, allowing us
to stay more consistent with the IDL files spec's provide.

* WebCore.xcodeproj/project.pbxproj:
Move JSWebGL2RenderingContextCustom to the GC/Wrapping only group.

* bindings/IDLTypes.h:
Add IDLWebGLAny, and simplify IDLIDBKey to not be templatized.

* bindings/js/JSDOMConvert.h:
(WebCore::Converter<IDLLong>::convert):
(WebCore::Converter<IDLFloat>::convert):
(WebCore::Converter<IDLUnrestrictedFloat>::convert):
(WebCore::Converter<IDLDouble>::convert):
(WebCore::Converter<IDLUnrestrictedDouble>::convert):
Add overloads of convert that convert a double to the numeric type.

(WebCore::Detail::GenericSequenceConverter::convert):
Add optimized specialization for JSArrays that does not use the iterator protocol.

(WebCore::Detail::NumericSequenceConverter::convert):
Add an even more optimized specialization for numeric types, which checks to see
if the array's butterfly has an optimized shape, and extracts the numeric values
from that.

(WebCore::Detail::SequenceConverter::convert):
(WebCore::Detail::SequenceConverter<IDLLong>::convert):
(WebCore::Detail::SequenceConverter<IDLFloat>::convert):
(WebCore::Detail::SequenceConverter<IDLUnrestrictedFloat>::convert):
(WebCore::Detail::SequenceConverter<IDLDouble>::convert):
(WebCore::Detail::SequenceConverter<IDLUnrestrictedDouble>::convert):
Select the right specialization based on type.

(WebCore::Converter<IDLSequence<T>>::convert):
(WebCore::Converter<IDLFrozenArray<T>>::convert):
Rename converter to be more accurate.

(WebCore::JSConverter<IDLWebGLAny>::convert):
Added. Calls through to converter in WebGLAny.h. We should replace this with a
call to the IDLUnion JSConverter when feasible.

* bindings/js/JSWebGL2RenderingContextCustom.cpp:
(WebCore::JSWebGL2RenderingContext::getInternalformatParameter): Deleted.
(WebCore::JSWebGL2RenderingContext::getQueryParameter): Deleted.
(WebCore::JSWebGL2RenderingContext::getSamplerParameter): Deleted.
(WebCore::JSWebGL2RenderingContext::getSyncParameter): Deleted.
(WebCore::JSWebGL2RenderingContext::getIndexedParameter): Deleted.
(WebCore::JSWebGL2RenderingContext::getActiveUniformBlockParameter): Deleted.
(WebCore::JSWebGL2RenderingContext::getActiveUniformBlockName): Deleted.
Remove unnecessary custom functions.

* bindings/js/JSWebGLRenderingContextBaseCustom.cpp:
(WebCore::objectParameter): Deleted.
(WebCore::JSWebGLRenderingContextBase::getAttachedShaders): Deleted.
(WebCore::JSWebGLRenderingContextBase::getBufferParameter): Deleted.
(WebCore::JSWebGLRenderingContextBase::getFramebufferAttachmentParameter): Deleted.
(WebCore::JSWebGLRenderingContextBase::getParameter): Deleted.
(WebCore::JSWebGLRenderingContextBase::getProgramParameter): Deleted.
(WebCore::JSWebGLRenderingContextBase::getRenderbufferParameter): Deleted.
(WebCore::JSWebGLRenderingContextBase::getShaderParameter): Deleted.
(WebCore::JSWebGLRenderingContextBase::getSupportedExtensions): Deleted.
(WebCore::JSWebGLRenderingContextBase::getTexParameter): Deleted.
(WebCore::JSWebGLRenderingContextBase::getUniform): Deleted.
(WebCore::JSWebGLRenderingContextBase::getVertexAttrib): Deleted.
(WebCore::toNumberVector): Deleted.
(WebCore::functionForUniform): Deleted.
(WebCore::dataFunctionf): Deleted.
(WebCore::dataFunctioni): Deleted.
(WebCore::dataFunctionMatrix): Deleted.
(WebCore::JSWebGLRenderingContextBase::uniform1fv): Deleted.
(WebCore::JSWebGLRenderingContextBase::uniform1iv): Deleted.
(WebCore::JSWebGLRenderingContextBase::uniform2fv): Deleted.
(WebCore::JSWebGLRenderingContextBase::uniform2iv): Deleted.
(WebCore::JSWebGLRenderingContextBase::uniform3fv): Deleted.
(WebCore::JSWebGLRenderingContextBase::uniform3iv): Deleted.
(WebCore::JSWebGLRenderingContextBase::uniform4fv): Deleted.
(WebCore::JSWebGLRenderingContextBase::uniform4iv): Deleted.
(WebCore::JSWebGLRenderingContextBase::uniformMatrix2fv): Deleted.
(WebCore::JSWebGLRenderingContextBase::uniformMatrix3fv): Deleted.
(WebCore::JSWebGLRenderingContextBase::uniformMatrix4fv): Deleted.
(WebCore::JSWebGLRenderingContextBase::vertexAttrib1fv): Deleted.
(WebCore::JSWebGLRenderingContextBase::vertexAttrib2fv): Deleted.
(WebCore::JSWebGLRenderingContextBase::vertexAttrib3fv): Deleted.
(WebCore::JSWebGLRenderingContextBase::vertexAttrib4fv): Deleted.
Remove unnecessary custom functions.

* bindings/scripts/CodeGenerator.pm:
(IsBuiltinType):
Remove specialization of IDBKey.

* bindings/scripts/CodeGeneratorJS.pm:
(GetBaseIDLType):
(GetIDLType):
(NativeToJSValueDOMConvertNeedsState):
(NativeToJSValueDOMConvertNeedsGlobalObject):
(NativeToJSValue):
Allow IDL files to override the type of signature using a new OverrideIDLType
extended attribute. Currently this only works conversion to JS, but could be
extended to support more.

* bindings/scripts/IDLAttributes.txt:
Add OverrideIDLType.

* html/canvas/WebGL2RenderingContext.idl:
Add OverrideIDLType=IDLWebGLAny annotations to remove the need for custom functions.

* html/canvas/WebGLAny.cpp:
(WebCore::convertToJSValue):
(WebCore::toJS): Deleted.
* html/canvas/WebGLAny.h:
Rename the toJS function to convertToJSValue to avoid a template specialization issue.

* html/canvas/WebGLRenderingContextBase.cpp:
(WebCore::WebGLRenderingContextBase::getAttachedShaders):
Update signature to match IDL, replacing bool return value plus out parameter with an
optional return value. Modernize the code a bit.

(WebCore::WebGLRenderingContextBase::uniform2f):
(WebCore::WebGLRenderingContextBase::uniform3f):
(WebCore::WebGLRenderingContextBase::uniform4f):
(WebCore::WebGLRenderingContextBase::uniform1i):
(WebCore::WebGLRenderingContextBase::uniform3i):
(WebCore::WebGLRenderingContextBase::uniform4i):
(WebCore::WebGLRenderingContextBase::uniform1fv):
(WebCore::WebGLRenderingContextBase::uniform2fv):
(WebCore::WebGLRenderingContextBase::uniform3fv):
(WebCore::WebGLRenderingContextBase::uniform4fv):
(WebCore::WebGLRenderingContextBase::uniform1iv):
(WebCore::WebGLRenderingContextBase::uniform2iv):
(WebCore::WebGLRenderingContextBase::uniform3iv):
(WebCore::WebGLRenderingContextBase::uniform4iv):
(WebCore::WebGLRenderingContextBase::uniformMatrix2fv):
(WebCore::WebGLRenderingContextBase::uniformMatrix3fv):
(WebCore::WebGLRenderingContextBase::uniformMatrix4fv):
(WebCore::WebGLRenderingContextBase::vertexAttrib4f):
(WebCore::WebGLRenderingContextBase::vertexAttrib1fv):
(WebCore::WebGLRenderingContextBase::vertexAttrib2fv):
(WebCore::WebGLRenderingContextBase::vertexAttrib3fv):
(WebCore::WebGLRenderingContextBase::vertexAttrib4fv):
(WebCore::WebGLRenderingContextBase::validateUniformParameters):
(WebCore::WebGLRenderingContextBase::validateUniformMatrixParameters):
(WebCore::WebGLRenderingContextBase::vertexAttribfvImpl):
* html/canvas/WebGLRenderingContextBase.h:
(WebCore::WebGLRenderingContextBase::TypedList::TypedList):
(WebCore::WebGLRenderingContextBase::TypedList::data):
(WebCore::WebGLRenderingContextBase::TypedList::length):
Re-order/group to match order in the IDL file. Remove overloads taking a pointer and
length, and replace it by having the main function take a specialized TypedList. TypedList
allows the bindings to pass the correct Variant type, but then the code to only think in
terms of the data() and length() functions.

* html/canvas/WebGLRenderingContextBase.idl:
Remove [Custom] annotations and re-order/group to match the spec. Update types of uniform*fv/iv
functions to take typedef'd unions.

LayoutTests:

* fast/canvas/webgl/gl-object-get-calls-expected.txt:
Update result for better error message.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210567 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoTemporarily add back the file removed in r210463 to unbreak internal bots.
rniwa@webkit.org [Tue, 10 Jan 2017 23:45:22 +0000 (23:45 +0000)]
Temporarily add back the file removed in r210463 to unbreak internal bots.

* Scripts/webkitpy/benchmark_runner/data/patches/Animometer.patch: Copied from Tools/Scripts/webkitpy/benchmark_runner/data/patches/Animometer.patch.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210566 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoStreamline the GC barrier slowpath
fpizlo@apple.com [Tue, 10 Jan 2017 23:41:55 +0000 (23:41 +0000)]
Streamline the GC barrier slowpath
https://bugs.webkit.org/show_bug.cgi?id=166878

Reviewed by Geoffrey Garen and Saam Barati.

This implements two optimizations to the barrier:

- Removes the write barrier buffer. This was just overhead.

- Teaches the slow path how to white an object that was black but unmarked, ensuring that
  we don't take slow path for this object again.

* JavaScriptCore.xcodeproj/project.pbxproj:
* dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::compileStoreBarrier):
* ftl/FTLLowerDFGToB3.cpp:
(JSC::FTL::DFG::LowerDFGToB3::emitStoreBarrier):
* heap/CellState.h:
* heap/Heap.cpp:
(JSC::Heap::Heap):
(JSC::Heap::markToFixpoint):
(JSC::Heap::addToRememberedSet):
(JSC::Heap::stopTheWorld):
(JSC::Heap::writeBarrierSlowPath):
(JSC::Heap::buildConstraintSet):
(JSC::Heap::flushWriteBarrierBuffer): Deleted.
* heap/Heap.h:
(JSC::Heap::writeBarrierBuffer): Deleted.
* heap/SlotVisitor.cpp:
(JSC::SlotVisitor::appendJSCellOrAuxiliary):
(JSC::SlotVisitor::setMarkedAndAppendToMarkStack):
(JSC::SlotVisitor::appendToMarkStack):
(JSC::SlotVisitor::visitChildren):
* heap/WriteBarrierBuffer.cpp: Removed.
* heap/WriteBarrierBuffer.h: Removed.
* jit/JITOperations.cpp:
* jit/JITOperations.h:
* runtime/JSCellInlines.h:
(JSC::JSCell::JSCell):
* runtime/StructureIDBlob.h:
(JSC::StructureIDBlob::StructureIDBlob):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210565 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago:active and :hover states may not be updated across slots
rniwa@webkit.org [Tue, 10 Jan 2017 23:34:51 +0000 (23:34 +0000)]
:active and :hover states may not be updated across slots
https://bugs.webkit.org/show_bug.cgi?id=166881
<rdar://problem/29944582>

Reviewed by Antti Koivisto.

Source/WebCore:

The bug was caused by updateHoverActiveState not updating :hover and :active states on elements
when nodes are assigned to slots because they were walking up the tree using parentOrShadowHostElement
and parentNode. Fixed the bug by using parentElementInComposedTree instead since :hover and :active
states need to be updated in accordance with the render tree, which is created from the "flat tree"
or the "composed tree" in WebKit's terminology (this is old terminology in the spec).

Tests: fast/shadow-dom/clear-active-state-in-shadow.html
       fast/shadow-dom/hover-over-nested-slotted-content.html

* dom/Document.cpp:
(WebCore::Document::updateHoverActiveState): Fixed the bug.
* dom/Node.cpp:
(WebCore::Node::parentElementInComposedTree): Added.
* dom/Node.h:

LayoutTests:

Added two regression tests; one for clearing :active state across a slot, and another one for clearing
a hover state on an ancestor of a slot to which a slot with the hovered element is assigned.

* fast/shadow-dom/clear-active-state-in-shadow-expected.html: Added.
* fast/shadow-dom/clear-active-state-in-shadow.html: Added.
* fast/shadow-dom/hover-over-nested-slotted-content-expected.html: Added.
* fast/shadow-dom/hover-over-nested-slotted-content.html: Added.
* platform/ios-simulator/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210564 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoProperty setters should not be called for bound arguments list entries.
mark.lam@apple.com [Tue, 10 Jan 2017 23:17:42 +0000 (23:17 +0000)]
Property setters should not be called for bound arguments list entries.
https://bugs.webkit.org/show_bug.cgi?id=165631

Reviewed by Filip Pizlo.

JSTests:

* stress/property-setters-should-not-be-called-for-bound-arguments-list-entries.js: Added.

Source/JavaScriptCore:

* builtins/FunctionPrototype.js:
(bind):
- use @putByValDirect to set the bound arguments so that we don't consult the
  prototype chain for setters.

* runtime/IntlDateTimeFormatPrototype.cpp:
(JSC::IntlDateTimeFormatPrototypeGetterFormat):
* runtime/IntlNumberFormatPrototype.cpp:
(JSC::IntlNumberFormatPrototypeGetterFormat):
- no need to create a bound arguments array because these bound functions binds
  no arguments according to the spec.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210563 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoMissing logging in IconLoader::startLoading
krollin@apple.com [Tue, 10 Jan 2017 23:09:44 +0000 (23:09 +0000)]
Missing logging in IconLoader::startLoading
https://bugs.webkit.org/show_bug.cgi?id=166904

Reviewed by Sam Weinig.

The LOG_ERROR in startLoading references 'resourceRequest', which has
been WTF::Moved as part of a preceding function call. As such, the
logging statement doesn't print out a possibly useful URL. Fix this by
making a copy of the URL before it's affected by the Move.

No new tests -- change only affects logging code.

* loader/icon/IconLoader.cpp:
(WebCore::IconLoader::startLoading):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210562 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoRecord/replay: fix range used for fuzzy matching
krollin@apple.com [Tue, 10 Jan 2017 22:31:06 +0000 (22:31 +0000)]
Record/replay: fix range used for fuzzy matching
https://bugs.webkit.org/show_bug.cgi?id=166041

Reviewed by Darin Adler and Alex Christensen.

Because of two bugs, the attempt to determine the range of URLs to
check as part of the process of fuzzy matching was failing. The intent
was to find the range of URLs that started with the same
<scheme://host:port> as a given URL. However, because of a reversed
test, the upper end of the range ended up being the "end()" iterator
of the entire collection of URLs. With that fixed, there was another
bug due to one URL being given as <scheme://host:port> and the other
given as <scheme://host:port/> (note the trailing slash). Both of
these issues are now fixed.

* NetworkProcess/capture/NetworkCaptureManager.cpp:
(WebKit::NetworkCapture::Manager::initialize):
(WebKit::NetworkCapture::Manager::findBestFuzzyMatch):
(WebKit::NetworkCapture::Manager::urlIdentifyingCommonDomain):
* NetworkProcess/capture/NetworkCaptureManager.h:
* NetworkProcess/capture/NetworkCaptureResource.cpp:
(WebKit::NetworkCapture::Resource::url):
(WebKit::NetworkCapture::Resource::urlIdentifyingCommonDomain):
(WebKit::NetworkCapture::Resource::baseURL): Deleted.
* NetworkProcess/capture/NetworkCaptureResource.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210561 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoImplement "proximity" scroll snapping
wenson_hsieh@apple.com [Tue, 10 Jan 2017 22:26:56 +0000 (22:26 +0000)]
Implement "proximity" scroll snapping
https://bugs.webkit.org/show_bug.cgi?id=135994
<rdar://problem/18162418>

Reviewed by Dean Jackson.

Source/WebCore:

Adds support for proximity scroll snapping. To do this, we introduce scroll offset ranges, a list of scroll
offset ranges that are plumbed alongside the list of scroll snap offsets. Similar to a snap offset, a snap
offset range contains scroll offsets on which scrolling is allowed to come to a rest within a scroll snapping
container. However, unlike normal snap offsets, scrolling may only come to rest within a snap offset range if
the predicted scroll offset already lies within the range. The new algorithm for selecting a target scroll snap
position given a destination offset is now:

-   If the scroll destination lies within a snap offset range, return the scroll destination
-   Otherwise, compute the nearest lower/upper snap offsets and lower/upper snap offset ranges
-   If scrolling ended with no velocity, return the nearest snap offset
-   If scrolling ended with positive velocity, choose the upper snap offset only if there is no snap offset
    range in between the scroll destination and the snap offset; else, choose the lower snap offset
-   If scrolling ended with negative velocity, choose the lower snap offset only if there is no snap offset
    range in between the scroll destination and the snap offset; else, choose the upper snap offset

The extra rule accounting for scroll offset ranges in between the scroll destination and a potential snap offset
handles the corner case where the user scrolls with momentum very lightly away from a snap offset, such that the
predicted scroll destination is still within proximity of the snap offset. In this case, the regular (mandatory
scroll snapping) behavior would be to snap to the next offset in the direction of momentum scrolling, but
instead, it is more intuitive to return to the original snap position.

We also move scrolling prediction logic into ScrollingMomentumCalculator and adopt the platform
_NSScrollingMomentumCalculator's destinationOrigin property when computing the predicted scroll destination.
Previously, we were simply multiplying by an empirically-derived constant to approximate the scroll destination,
but now that we are supporting proximity scroll snapping, we need more exact scroll destinaton prediction in
order to make sure that scrolling to a snap offset range feels natural.

Tests: tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-then-proximity.html
       tiled-drawing/scrolling/scroll-snap/scroll-snap-proximity-mainframe.html
       tiled-drawing/scrolling/scroll-snap/scroll-snap-proximity-overflow.html

* WebCore.xcodeproj/project.pbxproj:
* page/scrolling/AsyncScrollingCoordinator.cpp:
(WebCore::setStateScrollingNodeSnapOffsetsAsFloat):
(WebCore::AsyncScrollingCoordinator::updateOverflowScrollingNode):
(WebCore::AsyncScrollingCoordinator::updateScrollSnapPropertiesWithFrameView):

Make boilerplate changes to plumb lists of horizontal and vertical snap offset ranges alongside the lists of
horizontal and vertical snap offsets.

* page/scrolling/AxisScrollSnapOffsets.cpp:
(WebCore::snapOffsetRangesToString):
(WebCore::indicesOfNearestSnapOffsetRanges):
(WebCore::indicesOfNearestSnapOffsets):
(WebCore::adjustAxisSnapOffsetsForScrollExtent):
(WebCore::computeAxisProximitySnapOffsetRanges):
(WebCore::updateSnapOffsetsForScrollableArea):
(WebCore::closestSnapOffset):

Adjust the snap offset selection algorithm to take snap offset ranges into account. See above for more details.
Additionally, augment snap offset update logic to emit snap offset ranges for proximity scroll snapping. To do
this, we run the following steps on the final list of processed snap offsets:
-   Compute the proximity distance, which (for now) is arbitrarily 0.3 * the length or width of the scroll snap
    port, depending on whether scroll snapping is taking place in the X or Y axis.
-   For each pair of adjacent snap offsets, if they are more than 2 * proximity distance away from each other,
    emit a snap offset range starting from (lower snap offset + proximity distance) and ending on (upper snap
    offset + proximity distance).

* page/scrolling/AxisScrollSnapOffsets.h:
(WebCore::closestSnapOffset): Deleted.
* page/scrolling/ScrollSnapOffsetsInfo.h:

Introduce ScrollSnapOffsetsInfo, a struct which contains data relevant to scroll snapping. This includes
vertical and horizontal snap offsets, as well as vertical and horizontal snap offset ranges. Snap offset ranges
consist of a vector of ranges of scroll offsets.

* page/scrolling/ScrollingCoordinator.h:
* page/scrolling/ScrollingMomentumCalculator.cpp:
(WebCore::projectedInertialScrollDistance):
(WebCore::ScrollingMomentumCalculator::ScrollingMomentumCalculator):
(WebCore::ScrollingMomentumCalculator::setRetargetedScrollOffset):
(WebCore::ScrollingMomentumCalculator::predictedDestinationOffset):
(WebCore::ScrollingMomentumCalculator::create):
(WebCore::ScrollingMomentumCalculator::setPlatformMomentumScrollingPredictionEnabled):
(WebCore::BasicScrollingMomentumCalculator::BasicScrollingMomentumCalculator):
(WebCore::BasicScrollingMomentumCalculator::linearlyInterpolatedOffsetAtProgress):
(WebCore::BasicScrollingMomentumCalculator::initializeInterpolationCoefficientsIfNecessary):
(WebCore::BasicScrollingMomentumCalculator::initializeSnapProgressCurve):
* page/scrolling/ScrollingMomentumCalculator.h:
(WebCore::ScrollingMomentumCalculator::retargetedScrollOffset):
(WebCore::ScrollingMomentumCalculator::retargetedScrollOffsetDidChange):

Currently, the ScrollingMomentumCalculator is responsible for taking an initial position, initial velocity, and
target position and animating the scroll offset from the initial to target position. Now, we refactor the
ScrollingMomentumCalculator interface to no longer take a target offset upon initialization, and instead compute
the predicted scroll destination given initial position and velocity; clients of the ScrollingMomentumCalculator
then use this predicted scroll destination to compute a retargeted scroll offset and then call
setRetargetedScrollOffset on the calculator, which sets up an animation curve to the new retargeted offset. This
allows both the AppKit-based scrolling momentum calculator and platform-invariant momentum calculator to be used
interchangeably, while still allowing them to compute a destination offset from initial parameters of the
scroll.

* page/scrolling/ScrollingStateScrollingNode.cpp:
(WebCore::ScrollingStateScrollingNode::ScrollingStateScrollingNode):
(WebCore::ScrollingStateScrollingNode::setHorizontalSnapOffsets):
(WebCore::ScrollingStateScrollingNode::setVerticalSnapOffsets):
(WebCore::ScrollingStateScrollingNode::setHorizontalSnapOffsetRanges):
(WebCore::ScrollingStateScrollingNode::setVerticalSnapOffsetRanges):
* page/scrolling/ScrollingStateScrollingNode.h:
(WebCore::ScrollingStateScrollingNode::horizontalSnapOffsets):
(WebCore::ScrollingStateScrollingNode::verticalSnapOffsets):
(WebCore::ScrollingStateScrollingNode::horizontalSnapOffsetRanges):
(WebCore::ScrollingStateScrollingNode::verticalSnapOffsetRanges):
* page/scrolling/ScrollingTreeScrollingNode.cpp:
(WebCore::ScrollingTreeScrollingNode::commitStateBeforeChildren):
(WebCore::ScrollingTreeScrollingNode::dumpProperties):
* page/scrolling/ScrollingTreeScrollingNode.h:
(WebCore::ScrollingTreeScrollingNode::horizontalSnapOffsets):
(WebCore::ScrollingTreeScrollingNode::verticalSnapOffsets):
(WebCore::ScrollingTreeScrollingNode::horizontalSnapOffsetRanges):
(WebCore::ScrollingTreeScrollingNode::verticalSnapOffsetRanges):

Add more boilerplate support for snap offset ranges.

* page/scrolling/mac/ScrollingMomentumCalculatorMac.h:
* page/scrolling/mac/ScrollingMomentumCalculatorMac.mm:
(WebCore::ScrollingMomentumCalculator::create):
(WebCore::ScrollingMomentumCalculator::setPlatformMomentumScrollingPredictionEnabled):
(WebCore::ScrollingMomentumCalculatorMac::ScrollingMomentumCalculatorMac):
(WebCore::ScrollingMomentumCalculatorMac::scrollOffsetAfterElapsedTime):
(WebCore::ScrollingMomentumCalculatorMac::predictedDestinationOffset):
(WebCore::ScrollingMomentumCalculatorMac::retargetedScrollOffsetDidChange):
(WebCore::ScrollingMomentumCalculatorMac::animationDuration):
(WebCore::ScrollingMomentumCalculatorMac::requiresMomentumScrolling):
(WebCore::ScrollingMomentumCalculatorMac::ensurePlatformMomentumCalculator):

Hook into AppKit momentum scroll offset prediction.

* page/scrolling/mac/ScrollingTreeFrameScrollingNodeMac.mm:
(WebCore::convertToLayoutUnits):
(WebCore::ScrollingTreeFrameScrollingNodeMac::commitStateBeforeChildren):
* platform/ScrollableArea.cpp:
(WebCore::ScrollableArea::ensureSnapOffsetsInfo):
(WebCore::ScrollableArea::horizontalSnapOffsets):
(WebCore::ScrollableArea::horizontalSnapOffsetRanges):
(WebCore::ScrollableArea::verticalSnapOffsetRanges):
(WebCore::ScrollableArea::verticalSnapOffsets):
(WebCore::ScrollableArea::setHorizontalSnapOffsets):
(WebCore::ScrollableArea::setVerticalSnapOffsets):
(WebCore::ScrollableArea::setHorizontalSnapOffsetRanges):
(WebCore::ScrollableArea::setVerticalSnapOffsetRanges):
(WebCore::ScrollableArea::clearHorizontalSnapOffsets):
(WebCore::ScrollableArea::clearVerticalSnapOffsets):
* platform/ScrollableArea.h:
(WebCore::ScrollableArea::horizontalSnapOffsets): Deleted.
(WebCore::ScrollableArea::verticalSnapOffsets): Deleted.
* platform/cocoa/ScrollController.h:
* platform/cocoa/ScrollController.mm:
(WebCore::ScrollController::processWheelEventForScrollSnap):

Fix an issue where initial scrolling velocity would be set to zero at the end of a drag gesture.

(WebCore::ScrollController::updateScrollSnapState):
(WebCore::ScrollController::updateScrollSnapPoints):
(WebCore::ScrollController::setNearestScrollSnapIndexForAxisAndOffset):
* platform/cocoa/ScrollSnapAnimatorState.h:
(WebCore::ScrollSnapAnimatorState::snapOffsetsForAxis):
(WebCore::ScrollSnapAnimatorState::snapOffsetRangesForAxis):
(WebCore::ScrollSnapAnimatorState::setSnapOffsetsAndPositionRangesForAxis):
(WebCore::ScrollSnapAnimatorState::setSnapOffsetsForAxis): Deleted.
* platform/cocoa/ScrollSnapAnimatorState.mm:
(WebCore::ScrollSnapAnimatorState::setupAnimationForState):
(WebCore::ScrollSnapAnimatorState::targetOffsetForStartOffset):
(WebCore::projectedInertialScrollDistance): Deleted.
* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::updateScrollCoordinatedLayer):
* testing/Internals.cpp:
(WebCore::Internals::setPlatformMomentumScrollingPredictionEnabled):

Add a new hook for layout tests to force scrolling momentum calculators to use the platform-invariant momentum
scrolling prediction heuristic instead of the platform-dependent one.

(WebCore::Internals::scrollSnapOffsets):
* testing/Internals.h:
* testing/Internals.idl:

Source/WebKit2:

Adds boilerplate support for plumbing lists of snap offset ranges from the web process to the UI process
alongside the list of snap offsets.

* Shared/Scrolling/RemoteScrollingCoordinatorTransaction.cpp:
(ArgumentCoder<ScrollingStateScrollingNode>::encode):
(ArgumentCoder<ScrollingStateScrollingNode>::decode):
* Shared/WebCoreArgumentCoders.cpp:
(IPC::ArgumentCoder<ScrollOffsetRange<float>>::encode):
(IPC::ArgumentCoder<ScrollOffsetRange<float>>::decode):
* Shared/WebCoreArgumentCoders.h:
* UIProcess/Scrolling/ios/ScrollingTreeOverflowScrollingNodeIOS.mm:
(-[WKOverflowScrollViewDelegate scrollViewWillEndDragging:withVelocity:targetContentOffset:]):
* UIProcess/ios/RemoteScrollingCoordinatorProxyIOS.mm:

Adjust mainframe proximity scroll snapping logic to not subtract out the top content inset when there is no
active snap offset (i.e. when snapping rests in a snap offset range). Attempting to subtract out the top inset
in this case caused the scroll offset to jump after ending a drag with no momentum in a snap offset range.

(WebKit::RemoteScrollingCoordinatorProxy::adjustTargetContentOffsetForSnapping):
(WebKit::RemoteScrollingCoordinatorProxy::shouldSnapForMainFrameScrolling):
(WebKit::RemoteScrollingCoordinatorProxy::closestSnapOffsetForMainFrameScrolling):

LayoutTests:

Adds 3 new layout tests for proximity scroll snapping. Also tweaks some existing tests that test scroll snapping
after scrolling with momentum to use the custom heuristic for predicting scroll destination instead of platform
momentum scrolling. This ensures that the results of our layout tests that depend on predicting momentum scroll
destination are consistent across runs.

* tiled-drawing/scrolling/latched-div-with-scroll-snap.html:
* tiled-drawing/scrolling/scroll-snap/scroll-snap-iframe.html:
* tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-2d-overflow.html:
* tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-borders.html:
* tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-hidden-scrollbars.html:
* tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-mainframe-horizontal.html:
* tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-mainframe-slow-horizontal.html:
* tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-mainframe-slow-vertical.html:
* tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-mainframe-vertical-then-horizontal.html:
* tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-mainframe-vertical.html:
* tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-overflow-stateless.html:
* tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-overflow.html:
* tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-padding.html:
* tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-rotated.html:

Force these tests to use platform-independent scrolling momentum prediction, by multiplying the last scroll
delta upon release by a constant factor.

* tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-then-proximity-expected.txt: Added.
* tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-then-proximity.html: Added.

Tests that after changing scroll-snap-type from mandatory to proximity, swiping downwards no longer snaps the
scroll offset to the second box, but instead leaves the scroll offset somewhere in the middle of the first box.

* tiled-drawing/scrolling/scroll-snap/scroll-snap-proximity-mainframe-expected.txt: Added.
* tiled-drawing/scrolling/scroll-snap/scroll-snap-proximity-mainframe.html: Added.

Tests that when scroll-snap-type is proximity in the mainframe, scrolling slightly downwards snaps the scroll
offset back up to the top; scrolling somewhere in the middle of the first box does not snap the scroll offset;
and scrolling near the end of the first box snaps the scroll offset to the second box.

* tiled-drawing/scrolling/scroll-snap/scroll-snap-proximity-overflow-expected.txt: Added.
* tiled-drawing/scrolling/scroll-snap/scroll-snap-proximity-overflow.html: Added.

Similar to scroll-snap-proximity-mainframe.html, except for overflow scrolling instead of the mainframe.

* tiled-drawing/scrolling/scroll-snap/scroll-snap-scrolling-jumps-to-top.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210560 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoMake Event.initEvent()'s first parameter mandatory
cdumez@apple.com [Tue, 10 Jan 2017 21:59:15 +0000 (21:59 +0000)]
Make Event.initEvent()'s first parameter mandatory
https://bugs.webkit.org/show_bug.cgi?id=166894

Reviewed by Alex Christensen.

LayoutTests/imported/w3c:

Rebaseline existing W3C test now that we're closing to the specification.

* web-platform-tests/dom/interfaces-expected.txt:

Source/WebCore:

Make Event.initEvent()'s first parameter mandatory. As per the
specification, all parameters should be mandatory and we are
currently dicussing the possibility of making them optional in
the specification. However, having the first parameter optional
makes little sense as it would set the event's type to the string
"undefined".

Test: fast/events/initEvent-parameters.html

* dom/Event.idl:

LayoutTests:

Add layout test coverage.

* fast/events/initEvent-parameters-expected.txt: Added.
* fast/events/initEvent-parameters.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210559 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoCalling async arrow function which is in a class's member function will cause error
gskachkov@gmail.com [Tue, 10 Jan 2017 21:13:33 +0000 (21:13 +0000)]
Calling async arrow function which is in a class's member function will cause error
https://bugs.webkit.org/show_bug.cgi?id=166879

Reviewed by Saam Barati.

JSTests:

* stress/async-arrow-functions-lexical-binding-in-class.js: Added.
(shouldBe):
(shouldBeAsync):
(BaseClass.prototype.baseClassValue):
(BaseClass.prototype.get property):
(BaseClass):
(runSomething):
(ChildClass.prototype.classValue):
(ChildClass.prototype.get classProperty):
(ChildClass.prototype.asyncValueExp):
(ChildClass.prototype.asyncValueBody):
(ChildClass.prototype.asyncThisPropExp):
(ChildClass.prototype.asyncThisPropBody):
(ChildClass.prototype.asyncThisPropInEvalExp):
(ChildClass.prototype.asyncThisPropInEvalBody):
(ChildClass.prototype.asyncThisValueExp):
(ChildClass.prototype.asyncThisValueBody):
(ChildClass.prototype.asyncThisValueInEvalExp):
(ChildClass.prototype.asyncThisValueInEvalBody):
(ChildClass):
(ChildClass2):
(ChildClass2.prototype.classValue):
(ChildClass2.prototype.get classProperty):
* stress/async-arrow-functions-lexical-super-binding.js:

Source/JavaScriptCore:

Current patch fixed loading 'super' in async arrow function. Errored appear becuase
super was loaded always nevertherless if it used in async arrow function or not, but bytecompiler
put to arrow function context only if it used within arrow function. So to fix this issue we need to
check if super was used in arrow function.

* bytecompiler/BytecodeGenerator.h:
* bytecompiler/NodesCodegen.cpp:
(JSC::FunctionNode::emitBytecode):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210558 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoUnreviewed, rolling out r210537.
ryanhaddad@apple.com [Tue, 10 Jan 2017 20:42:25 +0000 (20:42 +0000)]
Unreviewed, rolling out r210537.
https://bugs.webkit.org/show_bug.cgi?id=166903

This change introduced JSC test failures (Requested by
ryanhaddad on #webkit).

Reverted changeset:

"Implement JSSourceCode to propagate SourceCode in module
pipeline"
https://bugs.webkit.org/show_bug.cgi?id=166861
http://trac.webkit.org/changeset/210537

Patch by Commit Queue <commit-queue@webkit.org> on 2017-01-10

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210557 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoREGRESSION(r189555): ImageDocument title no longer includes the size of the image
cdumez@apple.com [Tue, 10 Jan 2017 20:12:18 +0000 (20:12 +0000)]
REGRESSION(r189555): ImageDocument title no longer includes the size of the image
https://bugs.webkit.org/show_bug.cgi?id=166876
<rdar://problem/29939970>

Reviewed by Ryosuke Niwa.

Add a code comment to clarify fix that was made in r210536.

* html/ImageDocument.cpp:
(WebCore::ImageDocument::createDocumentStructure):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210556 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoAdd support for MediaKeys.generateRequest().
jer.noble@apple.com [Tue, 10 Jan 2017 20:09:21 +0000 (20:09 +0000)]
Add support for MediaKeys.generateRequest().
https://bugs.webkit.org/show_bug.cgi?id=166880

Reviewed by Xabier Rodriguez-Calvar.

Source/WebCore:

Test: media/encrypted-media/mock-MediaKeySession-generateRequest.html

Implement MediaKeySession::generateRequest(). This requires some additions
to CDM, CDMPrivate, and CDMInstance to support platform adoption and to
implement mock support for testing. Additionally, add an InitDataRegistry
object which can sanitize and extract key data from various initData types.
(At the moment, only the "keyids" initData type is implemented.)

Drive-by fixes: Change all instances of initDataType from String -> AtomicString.
Add "const" flags to all CDMPrivate methods that might be called from a const
reference.

* CMakeLists.txt:
* Modules/encryptedmedia/CDM.cpp:
(WebCore::CDM::supportsInitDataType):
(WebCore::CDM::sanitizeInitData):
(WebCore::CDM::supportsInitData):
* Modules/encryptedmedia/CDM.h:
* Modules/encryptedmedia/CDMInstance.h:
* Modules/encryptedmedia/CDMPrivate.h:
* Modules/encryptedmedia/InitDataRegistry.cpp: Added.
(WebCore::extractKeyIDsKeyids):
(WebCore::sanitizeKeyids):
(WebCore::sanitizeCenc):
(WebCore::extractKeyIDsCenc):
(WebCore::sanitizeWebM):
(WebCore::extractKeyIDsWebM):
(WebCore::InitDataRegistry::shared):
(WebCore::InitDataRegistry::InitDataRegistry):
(WebCore::InitDataRegistry::sanitizeInitData):
(WebCore::InitDataRegistry::extractKeyIDs):
(WebCore::InitDataRegistry::registerInitDataType):
* Modules/encryptedmedia/InitDataRegistry.h:
* Modules/encryptedmedia/MediaKeySession.cpp:
(WebCore::messageEventName):
(WebCore::MediaKeySession::MediaKeySession):
(WebCore::MediaKeySession::generateRequest):
(WebCore::MediaKeySession::enqueueMessage):
* Modules/encryptedmedia/MediaKeySession.h:
* Modules/encryptedmedia/MediaKeys.cpp:
(WebCore::MediaKeys::setServerCertificate):
* WebCore.xcodeproj/project.pbxproj:
* testing/MockCDMFactory.cpp:
(WebCore::MockCDMFactory::addKeysToSessionWithID):
(WebCore::MockCDMFactory::setSupportedDataTypes):
(WebCore::MockCDM::supportsInitDataType):
(WebCore::MockCDM::supportsConfiguration):
(WebCore::MockCDM::supportsConfigurationWithRestrictions):
(WebCore::MockCDM::supportsSessionTypeWithConfiguration):
(WebCore::MockCDM::supportsRobustness):
(WebCore::MockCDM::distinctiveIdentifiersRequirement):
(WebCore::MockCDM::persistentStateRequirement):
(WebCore::MockCDM::distinctiveIdentifiersAreUniquePerOriginAndClearable):
(WebCore::MockCDM::supportsInitData):
(WebCore::MockCDMInstance::setServerCertificate):
(WebCore::MockCDMInstance::requestLicense):
* testing/MockCDMFactory.h:
(WebCore::MockCDMFactory::supportedDataTypes):
(WebCore::MockCDMFactory::hasSessionWithID):
(WebCore::MockCDMFactory::removeSessionWithID):
(WebCore::MockCDMFactory::setSupportedDataTypes): Deleted.

LayoutTests:

Add shouldResolve() and shouldReject() to video-test.js and update existing
tests to use these new functions.

* media/encrypted-media/mock-MediaKeys-createSession-expected.txt: Added.
* media/encrypted-media/mock-MediaKeys-createSession.html: Added.
* media/encrypted-media/mock-MediaKeys-setServerCertificate.html:
* media/encrypted-media/mock-MediaKeySession-generateRequest.html:
* media/encrypted-media/mock-MediaKeySystemAccess.html:
* media/encrypted-media/mock-navigator-requestMediaKeySystemAccess.html:
* media/video-test.js:
(shouldResolve):
(shouldReject):
* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210555 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoUnreviewed, rolling out r210540.
commit-queue@webkit.org [Tue, 10 Jan 2017 19:13:06 +0000 (19:13 +0000)]
Unreviewed, rolling out r210540.
https://bugs.webkit.org/show_bug.cgi?id=166896

too crude for non-WebCore clients (Requested by kling on
#webkit).

Reverted changeset:

"Crash when GC heap grows way too large."
https://bugs.webkit.org/show_bug.cgi?id=166875
http://trac.webkit.org/changeset/210540

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210554 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoJSArray has some object scanning races
fpizlo@apple.com [Tue, 10 Jan 2017 18:44:45 +0000 (18:44 +0000)]
JSArray has some object scanning races
https://bugs.webkit.org/show_bug.cgi?id=166874

Reviewed by Mark Lam.

This fixes two separate bugs, both of which I detected by running
array-splice-contiguous.js in extreme anger:

1) Some of the paths of shifting and unshifting were not grabbing the internal cell
   lock. This was causing the array storage scan to crash, even though it was well
   synchronized (the scan does hold the lock). The fix is just to hold the lock anywhere
   that memmoves the innards of the butterfly.

2) Out of line property scanning was synchronized using double collect snapshot. Array
   storage scanning was synchronized using locks. But what if array storage
   transformations messed up the out of line properties? It turns out that we actually
   need to hoist the array storage scanner's locking up into the double collect
   snapshot.

I don't know how to write a test that does any better of a job of catching this than
array-splice-contiguous.js.

* heap/DeferGC.h: Make DisallowGC usable even if NDEBUG.
* runtime/JSArray.cpp:
(JSC::JSArray::unshiftCountSlowCase):
(JSC::JSArray::shiftCountWithArrayStorage):
(JSC::JSArray::unshiftCountWithArrayStorage):
* runtime/JSObject.cpp:
(JSC::JSObject::visitButterflyImpl):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210553 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoAdd support for MediaKeys.createSession().
jer.noble@apple.com [Tue, 10 Jan 2017 18:41:31 +0000 (18:41 +0000)]
Add support for MediaKeys.createSession().
https://bugs.webkit.org/show_bug.cgi?id=166796

Reviewed by Eric Carlson.

Source/WebCore:

Test: media/encrypted-media/mock-MediaKeys-createSession.html

Implement MediaKeys::createSession(). This requires some additions to
CDM, CDMPrivate, and CDMInstance to support platform adoption and to
implement mock support for testing. Because the CDMInstance will now
be shared between MediaKeys and MediaKeySession, it cannot be a unique_ptr.
Make CDMInterface ref-counted and store it as a Ref.

* Modules/encryptedmedia/CDM.cpp:
(WebCore::CDM::createInstance):
(WebCore::CDM::supportsSessions):
* Modules/encryptedmedia/CDM.h:
* Modules/encryptedmedia/CDMInstance.h:
* Modules/encryptedmedia/CDMPrivate.h:
* Modules/encryptedmedia/MediaKeySession.cpp:
(WebCore::MediaKeySession::create):
(WebCore::MediaKeySession::MediaKeySession):
(WebCore::MediaKeySession::sessionId):
(WebCore::MediaKeySession::expiration):
(WebCore::MediaKeySession::keyStatuses):
* Modules/encryptedmedia/MediaKeySession.h:
* Modules/encryptedmedia/MediaKeySessionType.idl:
* Modules/encryptedmedia/MediaKeySystemAccess.cpp:
(WebCore::MediaKeySystemAccess::createMediaKeys):
* Modules/encryptedmedia/MediaKeys.cpp:
(WebCore::MediaKeys::MediaKeys):
(WebCore::MediaKeys::createSession):
* Modules/encryptedmedia/MediaKeys.h:
(WebCore::MediaKeys::create):
* Modules/encryptedmedia/MediaKeys.idl:

LayoutTests:

* media/encrypted-media/mock-MediaKeys-createSession-expected.txt: Added.
* media/encrypted-media/mock-MediaKeys-createSession.html: Added.
* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210552 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agosvn-apply failed to apply a patch (deleting file with svn property)
jbedard@apple.com [Tue, 10 Jan 2017 18:05:07 +0000 (18:05 +0000)]
svn-apply failed to apply a patch (deleting file with svn property)
https://bugs.webkit.org/show_bug.cgi?id=165953

Reviewed by Daniel Bates.

* Scripts/VCSUtils.pm:
(parseSvnPropertyValue): The beginning of a diff will cause an early exit in the case of no empty newlines.
* Scripts/webkitperl/VCSUtils_unittest/parseSvnProperty.pl: Add tests without break between diffs for properties.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210551 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoRemove pointerLockElement from DOMDocumentPrivate.h
rniwa@webkit.org [Tue, 10 Jan 2017 17:54:06 +0000 (17:54 +0000)]
Remove pointerLockElement from DOMDocumentPrivate.h
https://bugs.webkit.org/show_bug.cgi?id=166882

Reviewed by Simon Fraser.

Removed pointerLockElement declaration that was supposed to be removed in r209637.

* DOM/DOMDocumentPrivate.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210550 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoAdd support for MediaKeys.setServerCertificate()
jer.noble@apple.com [Tue, 10 Jan 2017 17:25:56 +0000 (17:25 +0000)]
Add support for MediaKeys.setServerCertificate()
https://bugs.webkit.org/show_bug.cgi?id=166772

Reviewed by Darin Adler.

Source/WebCore:

Test: media/encrypted-media/mock-MediaKeys-setServerCertificate.html

Implement MediaKeys::setServerCertificate(). This requires some additions
to CDM, CDMPrivate, and CDMInstance to support platform adoption and to
implement mock support for testing.

* Modules/encryptedmedia/CDM.cpp:
(WebCore::CDM::supportsServerCertificates):
* Modules/encryptedmedia/CDM.h:
* Modules/encryptedmedia/CDMInstance.h:
* Modules/encryptedmedia/CDMPrivate.h:
* Modules/encryptedmedia/MediaKeys.cpp:
(WebCore::MediaKeys::setServerCertificate):
* Modules/encryptedmedia/MediaKeys.h:
* testing/MockCDMFactory.cpp:
(WebCore::MockCDM::supportsServerCertificates):
(WebCore::MockCDMInstance::setServerCertificate):
* testing/MockCDMFactory.h:
(WebCore::MockCDMFactory::supportsServerCertificates):
(WebCore::MockCDMFactory::setSupportsServerCertificates):
* testing/MockCDMFactory.idl:

LayoutTests:

* media/encrypted-media/mock-MediaKeys-setServerCertificate-expected.txt: Added.
* media/encrypted-media/mock-MediaKeys-setServerCertificate.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210549 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoUpdate flakiness dashboard configuration.
ryanhaddad@apple.com [Tue, 10 Jan 2017 17:18:51 +0000 (17:18 +0000)]
Update flakiness dashboard configuration.
https://bugs.webkit.org/show_bug.cgi?id=166869

Reviewed by Alexey Proskuryakov.

* TestResultServer/static-dashboards/builders.jsonp:
* TestResultServer/static-dashboards/flakiness_dashboard.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210548 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[GTK] Stop honoring the environment variable USE_NATIVE_XDISPLAY
clopez@igalia.com [Tue, 10 Jan 2017 17:07:33 +0000 (17:07 +0000)]
[GTK] Stop honoring the environment variable USE_NATIVE_XDISPLAY
https://bugs.webkit.org/show_bug.cgi?id=165232

Unreviewed follow-up patch after r210539 and r210523.

Instead of using that environment variable, the argument
--display-server=xorg should be passed now.

The GTK+ performance bot was the only bot using this, and is
already passing --display-server=xorg after r210539.

* Scripts/webkitpy/port/gtk.py:
(GtkPort._driver_class):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210547 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoCachedScript cloning does not clone encodedSize
commit-queue@webkit.org [Tue, 10 Jan 2017 15:07:46 +0000 (15:07 +0000)]
CachedScript cloning does not clone encodedSize
https://bugs.webkit.org/show_bug.cgi?id=166865

Patch by Youenn Fablet <youenn@apple.com> on 2017-01-10
Reviewed by Darin Adler.

Source/WebCore:

Covered by updated test.

* loader/cache/CachedResource.cpp:
(WebCore::CachedResource::setBodyDataFrom): set encoded size based on being cloned resource.
* loader/cache/CachedScript.cpp:
(WebCore::CachedScript::script):
(WebCore::CachedScript::setBodyDataFrom): Making use of CachedResource::setBodyDataFrom for complete cloning.

LayoutTests:

* http/tests/security/cross-origin-cached-scripts-expected.txt:
* http/tests/security/cross-origin-cached-scripts.html:
* http/tests/security/resources/notify-loaded.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210546 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoThreadedCoordinatedLayerTreeHost::renderNextFrame() should short-cut to layer flushing
zandobersek@gmail.com [Tue, 10 Jan 2017 14:51:02 +0000 (14:51 +0000)]
ThreadedCoordinatedLayerTreeHost::renderNextFrame() should short-cut to layer flushing
https://bugs.webkit.org/show_bug.cgi?id=157614

Reviewed by Carlos Garcia Campos.

CoordinatedLayerTreeHost prevents any layer flushes while a previously-commited scene
state is still being rendered on the composition thread. renderNextFrame() is called
once that is complete, and a new layer flush is scheduled.

This change improves the whole ordeal by immediately performing the layer flush only if
it was requested during the time we were waiting on the renderer (i.e. when the latest
scene state was being composited), instead of scheduling it unconditionally.
m_scheduledWhileWaitingForRenderer member variable is added to track that occurrence.

* WebProcess/WebPage/CoordinatedGraphics/CoordinatedLayerTreeHost.cpp:
(WebKit::CoordinatedLayerTreeHost::scheduleLayerFlush):
(WebKit::CoordinatedLayerTreeHost::renderNextFrame):
* WebProcess/WebPage/CoordinatedGraphics/CoordinatedLayerTreeHost.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210545 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoUnreviewed. Remove the unused gReadyStateTimerInterval global from
zandobersek@gmail.com [Tue, 10 Jan 2017 14:49:56 +0000 (14:49 +0000)]
Unreviewed. Remove the unused gReadyStateTimerInterval global from
the MediaPlayerPrivateGStreamerMSE implementation file as it's only
springing up a compiler warning.

* platform/graphics/gstreamer/mse/MediaPlayerPrivateGStreamerMSE.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210544 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[CoordinatedGraphics] TextureMapperPlatformLayerProxy is constantly getting activated
zandobersek@gmail.com [Tue, 10 Jan 2017 12:32:17 +0000 (12:32 +0000)]
[CoordinatedGraphics] TextureMapperPlatformLayerProxy is constantly getting activated
https://bugs.webkit.org/show_bug.cgi?id=166856

Reviewed by Michael Catanzaro.

When CoordinatedGraphicsLayer::setContentsNeedsDisplay() is called, currently with
the threaded CoordinatedGraphics mode the platform layer is marked for synchronizing.
This means that at flush-time this layer swaps the buffer of the wrapped platform
layer and updates the layer state to indicate that the underlying
TextureMapperPlatformLayerProxy object has to be activated on the composition thread.

This current logic results in the TextureMapperPlatformLayerProxy object being
reactivated for every display request that arrives via setContentsNeedsDisplay().
This is pretty wasteful since this activation heap-allocates a RunLoop::Timer object
which internally also locks up the GLib main context to attach a new source, and
that's only after the source on the previous RunLoop::Timer was detached before
that RunLoop::Timer was destroyed.

To avoid this, CoordinatedGraphicsLayer::setContentsNeedsDisplay() sets a new flag,
m_shouldUpdatePlatformLayer. During flush, this flag is checked in
CoordinatedGraphicsLayer::updatePlatformLayer(), with the buffer swap performed in
case it's set. CoordinatedGraphicsLayer::syncPlatformLayer() now only sets the
platform layer proxy object for that layer's state.

* platform/graphics/texmap/coordinated/CoordinatedGraphicsLayer.cpp:
(WebCore::CoordinatedGraphicsLayer::CoordinatedGraphicsLayer):
(WebCore::CoordinatedGraphicsLayer::setContentsNeedsDisplay):
(WebCore::CoordinatedGraphicsLayer::syncPlatformLayer):
(WebCore::CoordinatedGraphicsLayer::updatePlatformLayer):
(WebCore::CoordinatedGraphicsLayer::flushCompositingStateForThisLayerOnly):
* platform/graphics/texmap/coordinated/CoordinatedGraphicsLayer.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210543 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[TexMap] Use WTF::Function in TextureMapperPlatformLayerProxy
zandobersek@gmail.com [Tue, 10 Jan 2017 12:29:36 +0000 (12:29 +0000)]
[TexMap] Use WTF::Function in TextureMapperPlatformLayerProxy
https://bugs.webkit.org/show_bug.cgi?id=166853

Reviewed by Michael Catanzaro.

Use WTF::Function in the TextureMapperPlatformLayerProxy, instead of
the std::function bloat.

* platform/graphics/texmap/TextureMapperPlatformLayerProxy.cpp:
(WebCore::TextureMapperPlatformLayerProxy::scheduleUpdateOnCompositorThread):
(WebCore::TextureMapperPlatformLayerProxy::compositorThreadUpdateTimerFired):
* platform/graphics/texmap/TextureMapperPlatformLayerProxy.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210542 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[TextureMapper] TextureMapperPlatformLayerProxy::swapBuffers() should hold the lock...
zandobersek@gmail.com [Tue, 10 Jan 2017 12:28:46 +0000 (12:28 +0000)]
[TextureMapper] TextureMapperPlatformLayerProxy::swapBuffers() should hold the lock throughout invocation
https://bugs.webkit.org/show_bug.cgi?id=166852

Reviewed by Michael Catanzaro.

* platform/graphics/texmap/TextureMapperPlatformLayerProxy.cpp:
(WebCore::TextureMapperPlatformLayerProxy::swapBuffer): Append to previous buffer
to the m_usedBuffers Vector while the object's lock is held. Everywhere else the
m_usedBuffers Vector is used under the lock, no reason to not do the same here.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210541 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoCrash when GC heap grows way too large.
akling@apple.com [Tue, 10 Jan 2017 12:16:35 +0000 (12:16 +0000)]
Crash when GC heap grows way too large.
<https://webkit.org/b/166875>
<rdar://problem/27896585>

Reviewed by Mark Lam.

Hard cap the JavaScript heap at 4GB of live objects (determined post-GC.)
If we go past this limit, crash with a recognizable signature.

* heap/Heap.cpp:
(JSC::Heap::didExceedHeapSizeLimit):
(JSC::Heap::updateAllocationLimits):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210540 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[GTK] Add a new release test bot for Wayland and pass --display-server=xorg to the...
clopez@igalia.com [Tue, 10 Jan 2017 11:03:12 +0000 (11:03 +0000)]
[GTK] Add a new release test bot for Wayland and pass --display-server=xorg to the GTK Perf bot.
https://bugs.webkit.org/show_bug.cgi?id=166872

Reviewed by Michael Catanzaro.

Add the new wayland release test bot as a test-only bot. It will download the built product from the release build-only bot.
Define the required "--display-server=driver" in the bot additionalArguments config entry.
Implement support for honoring additionalArguments on the RunGtkAPITests and RunAndUploadPerfTests steps.
The step RunWebKitTests already has support for it.

* BuildSlaveSupport/build.webkit.org-config/config.json:
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
(RunGtkAPITests.start):
(RunAndUploadPerfTests.start):
* BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210539 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoAdd an iOS test for tapping on a text node assigned to a slot
rniwa@webkit.org [Tue, 10 Jan 2017 07:02:20 +0000 (07:02 +0000)]
Add an iOS test for tapping on a text node assigned to a slot
https://bugs.webkit.org/show_bug.cgi?id=166877

Reviewed by Andy Estes.

Add a test for tapping on a text node assigned to a slot.
touchstart event must be fired on a parent of the slot with a touch event handler.

* TestExpectations:
* fast/shadow-dom/touch-event-on-text-assigned-to-slot-expected.txt: Added.
* fast/shadow-dom/touch-event-on-text-assigned-to-slot.html: Added.
* resources/ui-helper.js:
(window.UIHelper.tapAt): Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210538 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoImplement JSSourceCode to propagate SourceCode in module pipeline
utatane.tea@gmail.com [Tue, 10 Jan 2017 05:58:19 +0000 (05:58 +0000)]
Implement JSSourceCode to propagate SourceCode in module pipeline
https://bugs.webkit.org/show_bug.cgi?id=166861

Reviewed by Saam Barati.

Source/JavaScriptCore:

Instead of propagating source code string, we propagate JSSourceCode
cell in the module pipeline. This allows us to attach a metadata
to the propagated source code string. In particular, it propagates
SourceOrigin through the module pipeline.

* CMakeLists.txt:
* JavaScriptCore.xcodeproj/project.pbxproj:
* builtins/ModuleLoaderPrototype.js:
(fulfillFetch):
(requestFetch):
* jsc.cpp:
(GlobalObject::moduleLoaderFetch):
* llint/LLIntData.cpp:
(JSC::LLInt::Data::performAssertions):
* llint/LowLevelInterpreter.asm:
* runtime/Completion.cpp:
(JSC::loadAndEvaluateModule):
(JSC::loadModule):
* runtime/JSModuleLoader.cpp:
(JSC::JSModuleLoader::provide):
* runtime/JSModuleLoader.h:
* runtime/JSSourceCode.cpp: Added.
(JSC::JSSourceCode::destroy):
* runtime/JSSourceCode.h: Added.
(JSC::JSSourceCode::createStructure):
(JSC::JSSourceCode::create):
(JSC::JSSourceCode::sourceCode):
(JSC::JSSourceCode::JSSourceCode):
* runtime/JSType.h:
* runtime/ModuleLoaderPrototype.cpp:
(JSC::moduleLoaderPrototypeParseModule):
* runtime/VM.cpp:
(JSC::VM::VM):
* runtime/VM.h:

Source/WebCore:

Now SourceOrigin is correctly propagated through the module pipeline.
So the error reported by http/tests/security/contentSecurityPolicy/module-eval-blocked-expected.html
is updated.

* ForwardingHeaders/runtime/JSSourceCode.h: Added.
* bindings/js/JSDOMPromise.h:
(WebCore::DeferredPromise::resolveWithCallback):
* bindings/js/ScriptModuleLoader.cpp:
(WebCore::ScriptModuleLoader::notifyFinished):

LayoutTests:

* http/tests/security/contentSecurityPolicy/module-eval-blocked-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210537 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoREGRESSION(r189555): ImageDocument title no longer includes the size of the image
cdumez@apple.com [Tue, 10 Jan 2017 05:04:36 +0000 (05:04 +0000)]
REGRESSION(r189555): ImageDocument title no longer includes the size of the image
https://bugs.webkit.org/show_bug.cgi?id=166876
<rdar://problem/29939970>

Reviewed by Tim Horton.

Source/WebCore:

ImageDocument title no longer included the size of the image after r189555.
This is because Document::setTitle() is a no-op if the document does not
have a <head> element. To address the issue, we now include a <head>
element in ImageDocuments so that their title element properly gets added
to it.

Test: fast/images/imageDocument-title.html

* html/ImageDocument.cpp:
(WebCore::ImageDocument::createDocumentStructure):

LayoutTests:

Add layout test coverage.

* fast/images/imageDocument-title-expected.txt: Added.
* fast/images/imageDocument-title.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210536 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoREGRESSION (r210522): ASSERTION FAILED: divot.offset >= divotStart.offset seen with...
utatane.tea@gmail.com [Tue, 10 Jan 2017 04:19:43 +0000 (04:19 +0000)]
REGRESSION (r210522): ASSERTION FAILED: divot.offset >= divotStart.offset seen with stress/import-basic.js and stress/import-from-eval.js
https://bugs.webkit.org/show_bug.cgi?id=166873

Reviewed by Saam Barati.

The divot should be the end of `import` token.

* parser/Parser.cpp:
(JSC::Parser<LexerType>::parseMemberExpression):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210535 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoASSERTION FAILED: newLogicalTop >= logicalTop in WebCore::RenderBlockFlow::getClearDelta
zalan@apple.com [Tue, 10 Jan 2017 02:17:19 +0000 (02:17 +0000)]
ASSERTION FAILED: newLogicalTop >= logicalTop in WebCore::RenderBlockFlow::getClearDelta
https://bugs.webkit.org/show_bug.cgi?id=151202
<rdar://problem/27711822>

Reviewed by Myles C. Maxfield.

Source/WebCore:

FindNextFloatLogicalBottomAdapter uses LayoutUnit::max() to flag m_nextLogicalBottom uninitialized.
However LayoutUnit::max() can also be a valid value for m_nextLogicalBottom.
FindNextFloatLogicalBottomAdapter::nextLogicalBottom() returns 0 instead of the actual value when
it sees m_nextLogicalBottom uninitialized. In certain cases, it confuses the caller and we end up
with a runaway loop.

Test: fast/block/float/assert-when-line-has-not-enough-space-left.html

* rendering/FloatingObjects.cpp:
(WebCore::FindNextFloatLogicalBottomAdapter::FindNextFloatLogicalBottomAdapter):
(WebCore::FindNextFloatLogicalBottomAdapter::highValue):
(WebCore::FindNextFloatLogicalBottomAdapter::nextLogicalBottom):
(WebCore::FindNextFloatLogicalBottomAdapter::nextShapeLogicalBottom):
(WebCore::FindNextFloatLogicalBottomAdapter::collectIfNeeded):

LayoutTests:

* fast/block/float/assert-when-line-has-not-enough-space-left-expected.txt: Added.
* fast/block/float/assert-when-line-has-not-enough-space-left.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210534 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoUnreviewed, rolling out r210531.
ryanhaddad@apple.com [Tue, 10 Jan 2017 01:28:53 +0000 (01:28 +0000)]
Unreviewed, rolling out r210531.

This change broke the 32-bit macOS build and caused LayoutTest
failures.

Reverted changeset:

"File scheme should not allow access of a resource on a
different volume."
https://bugs.webkit.org/show_bug.cgi?id=158552
http://trac.webkit.org/changeset/210531

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210533 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[iOS] Drop VNodeTracker
cdumez@apple.com [Tue, 10 Jan 2017 00:46:04 +0000 (00:46 +0000)]
[iOS] Drop VNodeTracker
https://bugs.webkit.org/show_bug.cgi?id=166868
<rdar://problem/29937587>

Reviewed by Andreas Kling.

Source/WebCore:

Drop VNodeTracker on iOS. This was introduced a while back to try and
address vnode exhaustion issues in long running page load tests.
However, there was no evidence that the VNodeTracker functionality
helped so there is little point in keeping the extra code complexity.

* WebCore.xcodeproj/project.pbxproj:
* platform/SharedBuffer.h:
* platform/VNodeTracker.cpp: Removed.
* platform/VNodeTracker.h: Removed.
* platform/cf/SharedBufferCF.cpp:
(WebCore::SharedBuffer::SharedBuffer):
* platform/cocoa/VNodeTrackerCocoa.cpp: Removed.

Source/WebKit2:

* WebProcess/cocoa/WebProcessCocoa.mm:
(WebKit::WebProcess::platformInitializeWebProcess):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210532 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoFile scheme should not allow access of a resource on a different volume.
bfulgham@apple.com [Tue, 10 Jan 2017 00:32:24 +0000 (00:32 +0000)]
File scheme should not allow access of a resource on a different volume.
https://bugs.webkit.org/show_bug.cgi?id=158552
<rdar://problem/15307582>

Reviewed by Alex Christensen.

Source/WebCore:

Revise SecurityOrigin to prevent files from one storage device (volume) from accessing content
on a different storage device (volume) unless universal access is enabled.

Pass the current file device as part of the NSURLRequest so that CFNetwork can reject loads
where the device changes in the midst of a load.

Also properly reflect that SecurityOrigin is never null by passing as a reference,
rather than as a pointer.

Tests: Tools/TestWebKitAPI/Tests/mac/CrossPartitionFileSchemeAccess.mm

* page/SecurityOrigin.cpp:
(WebCore::SecurityOrigin::canAccess): Pass argument as reference.
(WebCore::SecurityOrigin::passesFileCheck): Add check that file URLs refer to files in
the same storage volume.
(WebCore::SecurityOrigin::canDisplay): Add check that files share the same volume.
(WebCore::SecurityOrigin::isSameSchemeHostPort): Pass argument as reference.
* page/SecurityOrigin.h:
* platform/FileSystem.cpp:
(WebCore::filesHaveSameVolume): Added.
* platform/FileSystem.h:
* platform/network/cocoa/ResourceRequestCocoa.mm:
(WebCore::ResourceRequest::doUpdatePlatformRequest): If loading a file URL, tell CFNetwork
the storage device at the time of the start of the load so we can trigger a failure if this
changes during the load operation.
* platform/posix/FileSystemPOSIX.cpp:
(WebCore::getFileDeviceId): Added.
* platform/win/FileSystemWin.cpp:
(WebCore::getFileDeviceId): Added.

Tools:

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Add new files.
* TestWebKitAPI/Tests/mac/CrossPartitionFileSchemeAccess.html: Added.
* TestWebKitAPI/Tests/mac/CrossPartitionFileSchemeAccess.mm: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210531 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoUnreviewed, fix cloop.
fpizlo@apple.com [Tue, 10 Jan 2017 00:15:11 +0000 (00:15 +0000)]
Unreviewed, fix cloop.

* dfg/DFGPlanInlines.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210530 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoSelection animation flashing on editable content when zooming.
enrica@apple.com [Mon, 9 Jan 2017 23:17:52 +0000 (23:17 +0000)]
Selection animation flashing on editable content when zooming.
https://bugs.webkit.org/show_bug.cgi?id=166863
rdar://problem/29931470

Reviewed by Tim Horton.

Implementing _allowAnimatedUpdateSelectionRectView and returning NO,
prevents the selection animation in UIKit.

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

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210525 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoUnindenting text inside a blockquote can result in the text being reordered
timothy_horton@apple.com [Mon, 9 Jan 2017 22:35:27 +0000 (22:35 +0000)]
Unindenting text inside a blockquote can result in the text being reordered
https://bugs.webkit.org/show_bug.cgi?id=166813

Reviewed by Darin Adler and Ryosuke Niwa.

Test: editing/execCommand/unindent-nested-blockquote-with-inner-div.html

* editing/IndentOutdentCommand.cpp:
(WebCore::IndentOutdentCommand::outdentParagraph):
Start splitting the tree at the beginning of the content being unindented,
not at the containing block flow element, which could e.g. contain other
elements (or even other blockquotes).

* editing/execCommand/unindent-nested-blockquote-with-inner-div-expected.txt: Added.
* editing/execCommand/unindent-nested-blockquote-with-inner-div.html: Added.

Old test output:
second<br><blockquote><div><blockquote>first</blockquote><br></div></blockquote>
New test output:
<blockquote><div><blockquote>first</blockquote><br></div></blockquote>second

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210524 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[GTK][Wayland] Allow running the layout tests under a native Wayland environment.
clopez@igalia.com [Mon, 9 Jan 2017 22:28:43 +0000 (22:28 +0000)]
[GTK][Wayland] Allow running the layout tests under a native Wayland environment.
https://bugs.webkit.org/show_bug.cgi?id=165232

Reviewed by Michael Catanzaro.

Add a new WaylandDriver to run the tests in the user current wayland session.

Also merge all the configuration options for selecting the display server inside one
switch --display-server=xvfb/xorg/wayland/weston (defaulting to xvfb).

This switch is supported on the scripts: run-gtk-tests, run-perf-tests and run-webkit-tests.

* Scripts/run-gtk-tests: Add missing logging.basicConfig() initialization for the error logger.
(TestRunner._create_driver):
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
(parse_args):
* Scripts/webkitpy/performance_tests/perftestsrunner.py:
(PerfTestsRunner._parse_args):
* Scripts/webkitpy/port/gtk.py:
(GtkPort.__init__):
(GtkPort._driver_class):
(GtkPort._search_paths):
* Scripts/webkitpy/port/waylanddriver.py: Added.
(WaylandDriver):
(WaylandDriver.check_driver):
(WaylandDriver._setup_environ_for_test):
(WaylandDriver._start):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210523 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[JSC] Prototype dynamic-import
utatane.tea@gmail.com [Mon, 9 Jan 2017 22:02:47 +0000 (22:02 +0000)]
[JSC] Prototype dynamic-import
https://bugs.webkit.org/show_bug.cgi?id=165724

Reviewed by Saam Barati.

JSTests:

* stress/import-basic.js: Added.
(async.async.load):
(async):
(catch):
* stress/import-from-eval.js: Added.
(async):
(catch):
* stress/import-syntax.js: Added.
(testSyntaxError):
* stress/import-tests/cocoa.js: Added.
(export.Cocoa):
(export.hello):
* stress/import-tests/multiple.js: Added.
(export.result):
* stress/import-tests/multiple2.js: Added.
(export.ok):
* stress/import-tests/should.js: Added.
(export.shouldBe):
(export.shouldThrow):
* stress/modules-syntax-error.js:

Source/JavaScriptCore:

In this patch, we implement stage3 dynamic-import proposal[1].
This patch adds a new special operator `import`. And by using it, we can import
the module dynamically from modules and scripts. Before this feature, the module
is always imported statically and before executing the modules, importing the modules
needs to be done. And especially, the module can only be imported from the module.
So the classic script cannot import and use the modules. This dynamic-import relaxes
the above restrictions.

The typical dynamic-import form is the following.

    import("...").then(function (namespace) { ... });

You can pass any AssignmentExpression for the import operator. So you can determine
the importing modules dynamically.

    import(value).then(function (namespace) { ... });

And previously the module import declaration is only allowed in the top level statements.
But this import operator is just an expression. So you can use it in the function.
And you can use it conditionally.

    async function go(cond)
    {
        if (cond)
            return import("...");
        return undefined;
    }
    await go(true);

Currently, this patch just implements this feature only for the JSC shell.
JSC module loader requires a new hook, `importModule`. And the JSC shell implements
this hook. So, for now, this dynamic-import is not available in the browser side.
If you write this `import` call, it always returns the rejected promise.

import is implemented like a special operator similar to `super`.
This is because import is context-sensitive. If you call the `import`, the module
key resolution is done based on the caller's running context.

For example, if you are running the script which filename is "./ok/hello.js", the module
key for the call`import("./resource/syntax.js")` becomes `"./ok/resource/syntax.js"`.
But if you write the completely same import form in the script "./error/hello.js", the
key becomes "./error/resource/syntax.js". So exposing this feature as the `import`
function is misleading: this function becomes caller's context-sensitive. That's why
dynamic-import is specified as a special operator.

To resolve the module key, we need the caller's context information like the filename of
the caller. This is provided by the SourceOrigin implemented in r210149.
In the JSC shell implementation, this SourceOrigin holds the filename of the caller. So
based on this implementation, the module loader resolve the module key.
In the near future, we will extend this SourceOrigin to hold more information needed for
the browser-side import implementation.

[1]: https://tc39.github.io/proposal-dynamic-import/

* builtins/ModuleLoaderPrototype.js:
(importModule):
* bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::emitGetTemplateObject):
(JSC::BytecodeGenerator::emitGetGlobalPrivate):
* bytecompiler/BytecodeGenerator.h:
* bytecompiler/NodesCodegen.cpp:
(JSC::ImportNode::emitBytecode):
* jsc.cpp:
(absolutePath):
(GlobalObject::moduleLoaderImportModule):
(functionRun):
(functionLoad):
(functionCheckSyntax):
(runWithScripts):
* parser/ASTBuilder.h:
(JSC::ASTBuilder::createImportExpr):
* parser/NodeConstructors.h:
(JSC::ImportNode::ImportNode):
* parser/Nodes.h:
(JSC::ExpressionNode::isImportNode):
* parser/Parser.cpp:
(JSC::Parser<LexerType>::parseMemberExpression):
* parser/SyntaxChecker.h:
(JSC::SyntaxChecker::createImportExpr):
* runtime/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init):
* runtime/JSGlobalObject.h:
* runtime/JSGlobalObjectFunctions.cpp:
(JSC::globalFuncImportModule):
* runtime/JSGlobalObjectFunctions.h:
* runtime/JSModuleLoader.cpp:
(JSC::JSModuleLoader::importModule):
(JSC::JSModuleLoader::getModuleNamespaceObject):
* runtime/JSModuleLoader.h:
* runtime/ModuleLoaderPrototype.cpp:
(JSC::moduleLoaderPrototypeGetModuleNamespaceObject):

Source/WebCore:

We do not set a handler for import for now.
So dynamic import feature is only enabled in the JSC shell right now.

* bindings/js/JSDOMWindowBase.cpp:
* bindings/js/JSWorkerGlobalScopeBase.cpp:

LayoutTests:

* sputnik/Conformance/07_Lexical_Conventions/7.5_Tokens/7.5.3_Future_Reserved_Words/S7.5.3_A1.16-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210522 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoMake the collector's fixpoint smart about scheduling work
fpizlo@apple.com [Mon, 9 Jan 2017 21:45:56 +0000 (21:45 +0000)]
Make the collector's fixpoint smart about scheduling work
https://bugs.webkit.org/show_bug.cgi?id=165910

Reviewed by Keith Miller.

Prior to this change, every time the GC would run any constraints in markToFixpoint, it
would run all of the constraints. It would always run them in the same order. That means
that so long as any one constraint was generating new work, we'd pay the price of all
constraints. This is usually OK because most constraints are cheap but it artificially
inflates the cost of slow constraints - especially ones that are expensive but usually
generate no new work.

This patch redoes how the GC runs constraints by applying ideas from data flow analysis.
The GC now builds a MarkingConstraintSet when it boots up, and this contains all of the
constraints as well as some meta-data about them. Now, markToFixpoint just calls into
MarkingConstraintSet to execute constraints. Because constraint execution and scheduling
need to be aware of each other, I rewrote markToFixpoint in such a way that it's more
obvious how the GC goes between constraint solving, marking with stopped mutator, and
marking with resumed mutator. This also changes the scheduler API in such a way that a
synchronous stop-the-world collection no longer needs to do fake stop/resume - instead we
just swap the space-time scheduler for the stop-the-world scheduler.

This is a big streamlining of the GC. This is a speed-up in GC-heavy tests because we
now execute most constraints exactly twice regardless of how many total fixpoint
iterations we do. Now, when we run out of marking work, the constraint solver will just
run the constraint that is most likely to generate new visiting work, and if it does
generate work, then the GC now goes back to marking. Before, it would run *all*
constraints and then go back to marking. The constraint solver is armed with three
information signals that it uses to sort the constraints in order of descending likelihood
to generate new marking work. Then it runs them in that order until it there is new
marking work. The signals are:

1) Whether the constraint is greyed by marking or execution. We call this the volatility
   of the constraint. For example, weak reference constraints have GreyedByMarking as
   their volatility because they are most likely to have something to say after we've done
   some marking. On the other hand, conservative roots have GreyedByExecution as their
   volatility because they will give new information anytime we let the mutator run. The
   constraint solver will only run GreyedByExecution constraints as roots and after the
   GreyedByMarking constraints go silent. This ensures that we don't try to scan
   conservative roots every time we need to re-run weak references and vice-versa.

   Another way to look at it is that the constraint solver tries to predict if the
   wavefront is advancing or retreating. The wavefront is almost certainly advancing so
   long as the mark stacks are non-empty or so long as at least one of the GreyedByMarking
   constraints is still producing work. Otherwise the wavefront is almost certainly
   retreating. It's most profitable to run GreyedByMarking constraints when the wavefront
   is advancing, and most profitable to run GreyedByExecution constraints when the
   wavefront is retreating.

   We use the predicted wavefront direction and the volatility of constraints as a
   first-order signal of constraint profitability.

2) How much visiting work was created the last time the constraint ran. The solver
   remembers the lastVisitCount, and uses it to predict how much work the constraint will
   generate next time. In practice this means we will keep re-running the one interesting
   constraint until it shuts up.

3) Optional work predictors for some constraints. The constraint that shuffles the mutator
   mark stack into the main SlotVisitor's mutator mark stack always knows exactly how much
   work it will create.

   The sum of (2) and (3) are used as a second-order signal of constraint profitability.

The constraint solver will always run all of the GreyedByExecution constraints at GC
start, since these double as the GC's roots. The constraint solver will always run all of
the GreyedByMarking constraints the first time that marking stalls. Other than that, the
solver will keep running constraints, sorted according to their likelihood to create work,
until either work is created or we run out of constraints to run. GC termination happens
when we run out of constraints to run.

This new infrastructure means that we have a much better chance of dealing with worst-case
DOM pathologies. If we can intelligently factor different evil DOM things into different
constraints with the right work predictions then this could reduce the cost of those DOM
things by a factor of N where N is the number of fixpoint iterations the GC typically
does. N is usually around 5-6 even for simple heaps.

My perf measurements say:

PLT3: 0.02% faster with 5.3% confidence.
JetStream: 0.15% faster with 17% confidence.
Speedometer: 0.58% faster with 82% confidence.

Here are the details from JetStream:

splay: 1.02173x faster with 0.996841 confidence
splay-latency: 1.0617x faster with 0.987462 confidence
towers.c: 1.01852x faster with 0.92128 confidence
crypto-md5: 1.06058x faster with 0.482363 confidence
score: 1.00152x faster with 0.16892 confidence

I think that Speedometer is legitimately benefiting from this change based on looking at
--logGC=true output. We are now spending less time reexecuting expensive constraints. I
think that JetStream/splay is also benefiting, because although the constraints it sees
are cheap, it spends 30% of its time in GC so even small improvements matter.

* CMakeLists.txt:
* JavaScriptCore.xcodeproj/project.pbxproj:
* dfg/DFGPlan.cpp:
(JSC::DFG::Plan::markCodeBlocks): Deleted.
(JSC::DFG::Plan::rememberCodeBlocks): Deleted.
* dfg/DFGPlan.h:
* dfg/DFGPlanInlines.h: Added.
(JSC::DFG::Plan::iterateCodeBlocksForGC):
* dfg/DFGWorklist.cpp:
(JSC::DFG::Worklist::markCodeBlocks): Deleted.
(JSC::DFG::Worklist::rememberCodeBlocks): Deleted.
(JSC::DFG::rememberCodeBlocks): Deleted.
* dfg/DFGWorklist.h:
* dfg/DFGWorklistInlines.h: Added.
(JSC::DFG::iterateCodeBlocksForGC):
(JSC::DFG::Worklist::iterateCodeBlocksForGC):
* heap/CodeBlockSet.cpp:
(JSC::CodeBlockSet::writeBarrierCurrentlyExecuting): Deleted.
* heap/CodeBlockSet.h:
(JSC::CodeBlockSet::iterate): Deleted.
* heap/CodeBlockSetInlines.h:
(JSC::CodeBlockSet::iterate):
(JSC::CodeBlockSet::iterateCurrentlyExecuting):
* heap/Heap.cpp:
(JSC::Heap::Heap):
(JSC::Heap::iterateExecutingAndCompilingCodeBlocks):
(JSC::Heap::iterateExecutingAndCompilingCodeBlocksWithoutHoldingLocks):
(JSC::Heap::assertSharedMarkStacksEmpty):
(JSC::Heap::markToFixpoint):
(JSC::Heap::endMarking):
(JSC::Heap::collectInThread):
(JSC::Heap::stopIfNecessarySlow):
(JSC::Heap::acquireAccessSlow):
(JSC::Heap::collectIfNecessaryOrDefer):
(JSC::Heap::buildConstraintSet):
(JSC::Heap::notifyIsSafeToCollect):
(JSC::Heap::ResumeTheWorldScope::ResumeTheWorldScope): Deleted.
(JSC::Heap::ResumeTheWorldScope::~ResumeTheWorldScope): Deleted.
(JSC::Heap::harvestWeakReferences): Deleted.
(JSC::Heap::visitConservativeRoots): Deleted.
(JSC::Heap::visitCompilerWorklistWeakReferences): Deleted.
* heap/Heap.h:
* heap/MarkingConstraint.cpp: Added.
(JSC::MarkingConstraint::MarkingConstraint):
(JSC::MarkingConstraint::~MarkingConstraint):
(JSC::MarkingConstraint::resetStats):
(JSC::MarkingConstraint::execute):
* heap/MarkingConstraint.h: Added.
(JSC::MarkingConstraint::index):
(JSC::MarkingConstraint::abbreviatedName):
(JSC::MarkingConstraint::name):
(JSC::MarkingConstraint::lastVisitCount):
(JSC::MarkingConstraint::quickWorkEstimate):
(JSC::MarkingConstraint::workEstimate):
(JSC::MarkingConstraint::volatility):
* heap/MarkingConstraintSet.cpp: Added.
(JSC::MarkingConstraintSet::ExecutionContext::ExecutionContext):
(JSC::MarkingConstraintSet::ExecutionContext::didVisitSomething):
(JSC::MarkingConstraintSet::ExecutionContext::shouldTimeOut):
(JSC::MarkingConstraintSet::ExecutionContext::drain):
(JSC::MarkingConstraintSet::ExecutionContext::didExecute):
(JSC::MarkingConstraintSet::ExecutionContext::execute):
(JSC::MarkingConstraintSet::MarkingConstraintSet):
(JSC::MarkingConstraintSet::~MarkingConstraintSet):
(JSC::MarkingConstraintSet::resetStats):
(JSC::MarkingConstraintSet::add):
(JSC::MarkingConstraintSet::executeBootstrap):
(JSC::MarkingConstraintSet::executeConvergence):
(JSC::MarkingConstraintSet::isWavefrontAdvancing):
(JSC::MarkingConstraintSet::executeConvergenceImpl):
(JSC::MarkingConstraintSet::executeAll):
* heap/MarkingConstraintSet.h: Added.
(JSC::MarkingConstraintSet::isWavefrontRetreating):
* heap/MutatorScheduler.cpp: Added.
(JSC::MutatorScheduler::MutatorScheduler):
(JSC::MutatorScheduler::~MutatorScheduler):
(JSC::MutatorScheduler::didStop):
(JSC::MutatorScheduler::willResume):
(JSC::MutatorScheduler::didExecuteConstraints):
(JSC::MutatorScheduler::log):
(JSC::MutatorScheduler::shouldStop):
(JSC::MutatorScheduler::shouldResume):
* heap/MutatorScheduler.h: Added.
* heap/OpaqueRootSet.h:
(JSC::OpaqueRootSet::add):
* heap/SlotVisitor.cpp:
(JSC::SlotVisitor::visitAsConstraint):
(JSC::SlotVisitor::drain):
(JSC::SlotVisitor::didReachTermination):
(JSC::SlotVisitor::hasWork):
(JSC::SlotVisitor::drainFromShared):
(JSC::SlotVisitor::drainInParallelPassively):
(JSC::SlotVisitor::addOpaqueRoot):
* heap/SlotVisitor.h:
(JSC::SlotVisitor::addToVisitCount):
* heap/SpaceTimeMutatorScheduler.cpp: Copied from Source/JavaScriptCore/heap/SpaceTimeScheduler.cpp.
(JSC::SpaceTimeMutatorScheduler::Snapshot::Snapshot):
(JSC::SpaceTimeMutatorScheduler::Snapshot::now):
(JSC::SpaceTimeMutatorScheduler::Snapshot::bytesAllocatedThisCycle):
(JSC::SpaceTimeMutatorScheduler::SpaceTimeMutatorScheduler):
(JSC::SpaceTimeMutatorScheduler::~SpaceTimeMutatorScheduler):
(JSC::SpaceTimeMutatorScheduler::state):
(JSC::SpaceTimeMutatorScheduler::beginCollection):
(JSC::SpaceTimeMutatorScheduler::didStop):
(JSC::SpaceTimeMutatorScheduler::willResume):
(JSC::SpaceTimeMutatorScheduler::didExecuteConstraints):
(JSC::SpaceTimeMutatorScheduler::timeToStop):
(JSC::SpaceTimeMutatorScheduler::timeToResume):
(JSC::SpaceTimeMutatorScheduler::log):
(JSC::SpaceTimeMutatorScheduler::endCollection):
(JSC::SpaceTimeMutatorScheduler::bytesAllocatedThisCycleImpl):
(JSC::SpaceTimeMutatorScheduler::bytesSinceBeginningOfCycle):
(JSC::SpaceTimeMutatorScheduler::maxHeadroom):
(JSC::SpaceTimeMutatorScheduler::headroomFullness):
(JSC::SpaceTimeMutatorScheduler::mutatorUtilization):
(JSC::SpaceTimeMutatorScheduler::collectorUtilization):
(JSC::SpaceTimeMutatorScheduler::elapsedInPeriod):
(JSC::SpaceTimeMutatorScheduler::phase):
(JSC::SpaceTimeMutatorScheduler::shouldBeResumed):
(JSC::SpaceTimeScheduler::Decision::targetMutatorUtilization): Deleted.
(JSC::SpaceTimeScheduler::Decision::targetCollectorUtilization): Deleted.
(JSC::SpaceTimeScheduler::Decision::elapsedInPeriod): Deleted.
(JSC::SpaceTimeScheduler::Decision::phase): Deleted.
(JSC::SpaceTimeScheduler::Decision::shouldBeResumed): Deleted.
(JSC::SpaceTimeScheduler::Decision::timeToResume): Deleted.
(JSC::SpaceTimeScheduler::Decision::timeToStop): Deleted.
(JSC::SpaceTimeScheduler::SpaceTimeScheduler): Deleted.
(JSC::SpaceTimeScheduler::snapPhase): Deleted.
(JSC::SpaceTimeScheduler::currentDecision): Deleted.
* heap/SpaceTimeMutatorScheduler.h: Copied from Source/JavaScriptCore/heap/SpaceTimeScheduler.h.
(JSC::SpaceTimeScheduler::Decision::operator bool): Deleted.
* heap/SpaceTimeScheduler.cpp: Removed.
* heap/SpaceTimeScheduler.h: Removed.
* heap/SynchronousStopTheWorldMutatorScheduler.cpp: Added.
(JSC::SynchronousStopTheWorldMutatorScheduler::SynchronousStopTheWorldMutatorScheduler):
(JSC::SynchronousStopTheWorldMutatorScheduler::~SynchronousStopTheWorldMutatorScheduler):
(JSC::SynchronousStopTheWorldMutatorScheduler::state):
(JSC::SynchronousStopTheWorldMutatorScheduler::beginCollection):
(JSC::SynchronousStopTheWorldMutatorScheduler::timeToStop):
(JSC::SynchronousStopTheWorldMutatorScheduler::timeToResume):
(JSC::SynchronousStopTheWorldMutatorScheduler::endCollection):
* heap/SynchronousStopTheWorldMutatorScheduler.h: Added.
* heap/VisitingTimeout.h: Added.
(JSC::VisitingTimeout::VisitingTimeout):
(JSC::VisitingTimeout::visitCount):
(JSC::VisitingTimeout::didVisitSomething):
(JSC::VisitingTimeout::shouldTimeOut):
* runtime/Options.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210521 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoUnreviewed, rolling out r210260.
commit-queue@webkit.org [Mon, 9 Jan 2017 21:31:50 +0000 (21:31 +0000)]
Unreviewed, rolling out r210260.
https://bugs.webkit.org/show_bug.cgi?id=166860

new colorpicker UI doesn't seem to work in trunk,
investigating offline (Requested by brrian on #webkit).

Reverted changeset:

"Web Inspector: color picker should feature an editable CSS
value"
https://bugs.webkit.org/show_bug.cgi?id=124356
http://trac.webkit.org/changeset/210260

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210520 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[QuickLook] Add a layout test for webkit.org/b/135651
aestes@apple.com [Mon, 9 Jan 2017 21:31:02 +0000 (21:31 +0000)]
[QuickLook] Add a layout test for webkit.org/b/135651
https://bugs.webkit.org/show_bug.cgi?id=166855

Reviewed by Andreas Kling.

Tools:

* DumpRenderTree/mac/FrameLoadDelegate.mm:
(-[FrameLoadDelegate webView:didFailProvisionalLoadWithError:forFrame:]): Added output for
errors with code WebKitErrorCannotShowURL.
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::didFailProvisionalLoadWithErrorForFrame): Made the error output
not specific to WebKit2 error codes.

LayoutTests:

* quicklook/nil-response-mime-type-expected.txt: Added.
* quicklook/nil-response-mime-type.html: Added.
* quicklook/resources/nil-response-mime-type.xls: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210519 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoUnreviewed, rolling out r210476.
commit-queue@webkit.org [Mon, 9 Jan 2017 21:09:04 +0000 (21:09 +0000)]
Unreviewed, rolling out r210476.
https://bugs.webkit.org/show_bug.cgi?id=166859

"4% JSBench regression" (Requested by keith_mi_ on #webkit).

Reverted changeset:

"Add a slice intrinsic to the DFG/FTL"
https://bugs.webkit.org/show_bug.cgi?id=166707
http://trac.webkit.org/changeset/210476

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210518 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoSupport webkit-test-runner key-value pairs in http tests in DumpRenderTree
commit-queue@webkit.org [Mon, 9 Jan 2017 19:34:48 +0000 (19:34 +0000)]
Support webkit-test-runner key-value pairs in http tests in DumpRenderTree
https://bugs.webkit.org/show_bug.cgi?id=166837

Patch by Antoine Quint <graouts@apple.com> on 2017-01-09
Reviewed by Alex Christensen.

We now correctly handle the --absolutePath argument in DumpRenderTree and use
the provided value when trying to determine the path to the test file. Failing
that, we use the computed URL or the raw path provided to cover all cases.

This was tested in the context of https://bugs.webkit.org/show_bug.cgi?id=165668
by applying this patch and seeing the number test failures dramatically reduced
with WK1 due to the option to turn modern media controls off for older media controls
tests.

* DumpRenderTree/DumpRenderTree.h:
* DumpRenderTree/DumpRenderTreeCommon.cpp:
(parseInputLine):
* DumpRenderTree/TestOptions.h:
* DumpRenderTree/TestOptions.mm:
(TestOptions::TestOptions):
* DumpRenderTree/mac/DumpRenderTree.mm:
(runTest):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210517 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoAdd a test to ensure that preventDefault on touchmove prevents scrolling
bdakin@apple.com [Mon, 9 Jan 2017 19:09:23 +0000 (19:09 +0000)]
Add a test to ensure that preventDefault on touchmove prevents scrolling
https://bugs.webkit.org/show_bug.cgi?id=166790

Reviewed by Simon Fraser.

* fast/events/touch/ios/iphone7/prevent-default-touchmove-prevents-scrolling-expected.txt: Added.
* fast/events/touch/ios/iphone7/prevent-default-touchmove-prevents-scrolling.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210516 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoRebaseline W3C test for ios-simulator after r210468.
ryanhaddad@apple.com [Mon, 9 Jan 2017 18:55:09 +0000 (18:55 +0000)]
Rebaseline W3C test for ios-simulator after r210468.
https://bugs.webkit.org/show_bug.cgi?id=166850

Unreviewed test gardening.

* platform/ios-simulator/imported/w3c/web-platform-tests/html/browsers/browsing-the-web/history-traversal/persisted-user-state-restoration/scroll-restoration-fragment-scrolling-cross-origin-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210515 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoMerging ThreadableLoader redundant options on filtering responses
commit-queue@webkit.org [Mon, 9 Jan 2017 18:26:58 +0000 (18:26 +0000)]
Merging ThreadableLoader redundant options on filtering responses

Merge OpaqueResponseBodyPolicy and ResponseFilteringPolicy ThreadableLoader options
https://bugs.webkit.org/show_bug.cgi?id=166843

Patch by Youenn Fablet <youennf@gmail.com> on 2017-01-09
Reviewed by Darin Adler.

Covered by existing tests.

Removing OpaqueResponseBodyPolicy option.
When ResponseFilteringPolicy is set to Enable, the body is filtered out in DocumentThreadableLoader.

* Modules/fetch/FetchLoader.cpp:
(WebCore::FetchLoader::start):
* loader/DocumentThreadableLoader.cpp:
(WebCore::DocumentThreadableLoader::didReceiveResponse): Implementing body filtering based on ResponseFilteringPolicy option.
* loader/ThreadableLoader.cpp:
(WebCore::ThreadableLoaderOptions::ThreadableLoaderOptions):
* loader/ThreadableLoader.h:
* loader/WorkerThreadableLoader.cpp:
(WebCore::LoaderTaskOptions::LoaderTaskOptions):
* workers/WorkerScriptLoader.cpp:
(WebCore::WorkerScriptLoader::loadSynchronously):
(WebCore::WorkerScriptLoader::loadAsynchronously):
* xml/XMLHttpRequest.cpp:
(WebCore::XMLHttpRequest::createRequest):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210514 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoprepare-ChangeLog: Correct help string for --style option
commit-queue@webkit.org [Mon, 9 Jan 2017 17:41:25 +0000 (17:41 +0000)]
prepare-ChangeLog: Correct help string for --style option
https://bugs.webkit.org/show_bug.cgi?id=166725

Patch by Milan Crha <mcrha@redhat.com> on 2017-01-09
Reviewed by Alexey Proskuryakov.

* Scripts/prepare-ChangeLog:
(main):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210512 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agohttps://bugs.webkit.org/show_bug.cgi?id=166839
lforschler@apple.com [Mon, 9 Jan 2017 17:39:14 +0000 (17:39 +0000)]
https://bugs.webkit.org/show_bug.cgi?id=166839
Add php7 support for macOS.
Tools:

Mirror the debian logic to look for a php7 module on disk, and return the appropriate apache config file.

Reviewed by Alexey Proskuryakov.

* Scripts/webkitpy/port/base.py:
(Port._is_darwin_php_version_7):
(Port._debian_php_version):
(Port):
(Port._darwin_php_version):
(Port._apache_config_file_name_for_platform):

LayoutTests:

This patch mirrors the solution that debian uses, by adding a php7 specific apache conf file.

Reviewed by Alexey Proskuryakov.

* http/conf/apache2.4-php7-httpd.conf: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210510 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoCreate new php7 config, preserving history via svn cp
lforschler@apple.com [Mon, 9 Jan 2017 17:33:15 +0000 (17:33 +0000)]
Create new php7 config, preserving history via svn cp

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210509 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoEvaluating window named element may return wrong result
dbates@webkit.org [Mon, 9 Jan 2017 17:19:46 +0000 (17:19 +0000)]
Evaluating window named element may return wrong result
https://bugs.webkit.org/show_bug.cgi?id=166792
<rdar://problem/29801059>

Reviewed by Chris Dumez.

* bindings/js/JSDOMWindowProperties.cpp:
(WebCore::jsDOMWindowPropertiesGetOwnPropertySlotNamedItemGetter):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210508 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoREGRESSION(r178955): Touching Settings.in doesn't cause JSInternalSettingsGenerated...
cdumez@apple.com [Mon, 9 Jan 2017 16:54:59 +0000 (16:54 +0000)]
REGRESSION(r178955): Touching Settings.in doesn't cause JSInternalSettingsGenerated.cpp to be updated on first build
https://bugs.webkit.org/show_bug.cgi?id=166814

Reviewed by Darin Adler.

Stop using .INTERMEDIATE in DerivedSources.make was it seems to break dependency
chains for some reason. We started using .INTERMEDIATE for scripts that generate
several targets when run a single time. It avoided running the script several
times (once per target) unnecessarily, and fixed issues with parallel builds.

To address the same issue without using .INTERMEDIATE, we now use pattern rules
(i.e. rules that have '%'). As per the documentation [1]:
"Pattern rules may have more than one target. Unlike normal rules, this does not
 act as many different rules with the same prerequisites and recipe. If a pattern
 rule has multiple targets, make knows that the rule’s recipe is responsible for
 making all of the targets. The recipe is executed only once to make all the
 targets".

 We are therefore solving the same problem as with .INTERMEDIATE. However, I have
 confirmed that using pattern rules does not break dependency chains. For example,
 modifying Settings.in re-generates both InternalSettingsGenerated.idl *and*
 JSInternalSettingsGenerated.cpp.

 [1] https://www.gnu.org/software/make/manual/html_node/Pattern-Intro.html#Pattern-Intro

* DerivedSources.make:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210507 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoUnreviewed, rolling out r210493 and r210495.
commit-queue@webkit.org [Mon, 9 Jan 2017 16:31:18 +0000 (16:31 +0000)]
Unreviewed, rolling out r210493 and r210495.
https://bugs.webkit.org/show_bug.cgi?id=166842

Causes makeprops.pl to run on every build (Requested by smfr
on #webkit).

Reverted changesets:

"Avoid triggering rebuilds for minor changes of
CSSProperties.json"
https://bugs.webkit.org/show_bug.cgi?id=166810
http://trac.webkit.org/changeset/210493

"Avoid triggering rebuilds for minor changes of
CSSProperties.json"
https://bugs.webkit.org/show_bug.cgi?id=166810
http://trac.webkit.org/changeset/210495

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210506 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoMarking inspector/debugger/breakpoint-action-log.html as flaky on macOS.
ryanhaddad@apple.com [Mon, 9 Jan 2017 16:02:39 +0000 (16:02 +0000)]
Marking inspector/debugger/breakpoint-action-log.html as flaky on macOS.
https://bugs.webkit.org/show_bug.cgi?id=165772

Unreviewed test gardening.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210505 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[GTK] Should support key and code properties on keyboard events
gns@gnome.org [Mon, 9 Jan 2017 14:14:57 +0000 (14:14 +0000)]
[GTK] Should support key and code properties on keyboard events
https://bugs.webkit.org/show_bug.cgi?id=166759

Reviewed by Carlos Garcia Campos.

Source/WebCore:

This change fixes a bunch of tests which are currently failing, such as:
- fast/events/arrow-keys-on-body.html
- fast/events/constructors/keyboard-event-constructor.html
- fast/events/key-events-in-input-button.html
- fast/events/key-events-in-input-text.html
- fast/events/keyboardevent-code.html
- fast/events/keyboardevent-key.html

* platform/PlatformKeyboardEvent.h:
* platform/gtk/PlatformKeyboardEventGtk.cpp:
(WebCore::PlatformKeyboardEvent::keyValueForGdkKeyCode): match web key value
to GDK key symbol.
(WebCore::PlatformKeyboardEvent::keyCodeForHardwareKeyCode): match web key code
to hardware code as reported by GDK.
(WebCore::PlatformKeyboardEvent::PlatformKeyboardEvent): add web key value and
code to the keyboard event upon creation.

Source/WebKit2:

* Shared/WebEvent.h:
* Shared/WebKeyboardEvent.cpp:
(WebKit::WebKeyboardEvent::WebKeyboardEvent): accept web key value and code on construction.
* Shared/gtk/WebEventFactory.cpp:
(WebKit::WebEventFactory::createWebKeyboardEvent): pass web key value and code on construction.

Source/WTF:

* wtf/FeatureDefines.h: enable key and code properties support for GTK.

Tools:

* WebKitTestRunner/gtk/EventSenderProxyGtk.cpp:
(WTR::getGDKKeySymForKeyRef): add missing modifier keys which were being reported as 'l' and 'r' only
on tests.

LayoutTests:

* platform/gtk/fast/events/keyboardevent-code-expected.txt: Added. Hardware
codes are different for Mac and Linux for a couple of keys, so we need
platform-specific results.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210504 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[Win] Some tests are flaky because certain DLLs are writing to stdout.
pvollan@apple.com [Mon, 9 Jan 2017 12:46:34 +0000 (12:46 +0000)]
[Win] Some tests are flaky because certain DLLs are writing to stdout.
https://bugs.webkit.org/show_bug.cgi?id=166760

Reviewed by Darin Adler.

Some tests are flaky because certain DLLs are writing to stdout, giving incorrect test results.
We can work around that by duplicating and redirecting stdout.

* DumpRenderTree/PixelDumpSupport.cpp:
(dumpWebViewAsPixelsAndCompareWithExpected):
(printPNG):
* DumpRenderTree/win/AccessibilityControllerWin.cpp:
(logEventProc):
* DumpRenderTree/win/DRTDesktopNotificationPresenter.cpp:
(DRTDesktopNotificationPresenter::showDesktopNotification):
(DRTDesktopNotificationPresenter::cancelDesktopNotification):
(DRTDesktopNotificationPresenter::requestNotificationPermission):
* DumpRenderTree/win/DumpRenderTree.cpp:
(dumpFrameScrollPosition):
(dumpHistoryItem):
(dumpBackForwardList):
(dump):
(runTest):
(main):
* DumpRenderTree/win/DumpRenderTreeWin.h:
* DumpRenderTree/win/EditingDelegate.cpp:
(EditingDelegate::shouldBeginEditingInDOMRange):
(EditingDelegate::shouldEndEditingInDOMRange):
(EditingDelegate::shouldInsertNode):
(EditingDelegate::shouldInsertText):
(EditingDelegate::shouldDeleteDOMRange):
(EditingDelegate::shouldChangeSelectedDOMRange):
(EditingDelegate::shouldApplyStyle):
(EditingDelegate::shouldChangeTypingStyle):
(EditingDelegate::doPlatformCommand):
(EditingDelegate::webViewDidBeginEditing):
(EditingDelegate::webViewDidChange):
(EditingDelegate::webViewDidEndEditing):
(EditingDelegate::webViewDidChangeTypingStyle):
(EditingDelegate::webViewDidChangeSelection):
* DumpRenderTree/win/FrameLoadDelegate.cpp:
(FrameLoadDelegate::didStartProvisionalLoadForFrame):
(FrameLoadDelegate::didReceiveServerRedirectForProvisionalLoadForFrame):
(FrameLoadDelegate::didChangeLocationWithinPageForFrame):
(FrameLoadDelegate::didFailProvisionalLoadWithError):
(FrameLoadDelegate::didCommitLoadForFrame):
(FrameLoadDelegate::didReceiveTitle):
(FrameLoadDelegate::didFinishLoadForFrame):
(FrameLoadDelegate::didFailLoadWithError):
(FrameLoadDelegate::willPerformClientRedirectToURL):
(FrameLoadDelegate::didCancelClientRedirectForFrame):
(FrameLoadDelegate::windowScriptObjectAvailable):
(FrameLoadDelegate::didFinishDocumentLoadForFrame):
(FrameLoadDelegate::didHandleOnloadEventsForFrame):
(FrameLoadDelegate::didDisplayInsecureContent):
(FrameLoadDelegate::didRunInsecureContent):
(FrameLoadDelegate::webViewProgressFinishedNotification):
* DumpRenderTree/win/HistoryDelegate.cpp:
(HistoryDelegate::didNavigateWithNavigationData):
(HistoryDelegate::didPerformClientRedirectFromURL):
(HistoryDelegate::didPerformServerRedirectFromURL):
(HistoryDelegate::updateHistoryTitle):
(HistoryDelegate::populateVisitedLinksForWebView):
* DumpRenderTree/win/PolicyDelegate.cpp:
(PolicyDelegate::decidePolicyForNavigationAction):
(PolicyDelegate::unableToImplementPolicyWithError):
* DumpRenderTree/win/ResourceLoadDelegate.cpp:
(ResourceLoadDelegate::willSendRequest):
(ResourceLoadDelegate::didReceiveAuthenticationChallenge):
(ResourceLoadDelegate::didReceiveResponse):
(ResourceLoadDelegate::didFinishLoadingFromDataSource):
(ResourceLoadDelegate::didFailLoadingWithError):
* DumpRenderTree/win/TestRunnerWin.cpp:
(TestRunner::addDisallowedURL):
(TestRunner::originsWithApplicationCache):
(TestRunner::setStorageDatabaseIdleInterval):
(TestRunner::closeIdleLocalStorageDatabases):
(TestRunner::copyDecodedHostName):
(TestRunner::copyEncodedHostName):
(TestRunner::numberOfPendingGeolocationPermissionRequests):
(TestRunner::isGeolocationProviderActive):
(TestRunner::pathToLocalResource):
(TestRunner::setMockDeviceOrientation):
(TestRunner::setMockGeolocationPosition):
(TestRunner::setMockGeolocationPositionUnavailableError):
(TestRunner::setSpatialNavigationEnabled):
(TestRunner::setAutomaticLinkDetectionEnabled):
(TestRunner::isCommandEnabled):
(TestRunner::authenticateSession):
(TestRunner::addChromeInputField):
(TestRunner::removeChromeInputField):
(TestRunner::focusWebView):
(TestRunner::grantWebNotificationPermission):
(TestRunner::denyWebNotificationPermission):
(TestRunner::removeAllWebNotificationPermissions):
(TestRunner::simulateWebNotificationClick):
(TestRunner::imageCountInGeneralPasteboard):
* DumpRenderTree/win/UIDelegate.cpp:
(UIDelegate::runJavaScriptAlertPanelWithMessage):
(UIDelegate::runJavaScriptConfirmPanelWithMessage):
(UIDelegate::runJavaScriptTextInputPanelWithPrompt):
(UIDelegate::runBeforeUnloadConfirmPanelWithMessage):
(UIDelegate::webViewAddMessageToConsole):
(UIDelegate::exceededDatabaseQuota):
(UIDelegate::setStatusText):
(UIDelegate::didPressMissingPluginButton):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210503 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoMove cache coders to WTF
antti@apple.com [Mon, 9 Jan 2017 12:43:49 +0000 (12:43 +0000)]
Move cache coders to WTF
https://bugs.webkit.org/show_bug.cgi?id=166825

Rubber-stamped by Sam Weinig.

Source/WebKit2:

* NetworkProcess/cache/NetworkCache.h:
* NetworkProcess/cache/NetworkCacheCoder.h: Removed.
* NetworkProcess/cache/NetworkCacheCoders.cpp:
(WebKit::NetworkCache::Coder<AtomicString>::encode): Deleted.
(WebKit::NetworkCache::Coder<AtomicString>::decode): Deleted.
(WebKit::NetworkCache::Coder<CString>::encode): Deleted.
(WebKit::NetworkCache::Coder<CString>::decode): Deleted.
(WebKit::NetworkCache::Coder<String>::encode): Deleted.
(WebKit::NetworkCache::decodeStringText): Deleted.
(WebKit::NetworkCache::Coder<String>::decode): Deleted.
(WebKit::NetworkCache::Coder<SHA1::Digest>::encode): Deleted.
(WebKit::NetworkCache::Coder<SHA1::Digest>::decode): Deleted.
(WebKit::NetworkCache::Coder<WebCore::HTTPHeaderMap>::encode): Deleted.
(WebKit::NetworkCache::Coder<WebCore::HTTPHeaderMap>::decode): Deleted.
* NetworkProcess/cache/NetworkCacheCoders.h:
(): Deleted.
(WebKit::NetworkCache::Coder<std::optional<T>>::encode): Deleted.
(WebKit::NetworkCache::Coder<std::optional<T>>::decode): Deleted.
* NetworkProcess/cache/NetworkCacheCodersCocoa.cpp:
(WebKit::NetworkCache::encodeCFData): Deleted.
(WebKit::NetworkCache::decodeCFData): Deleted.
(WebKit::NetworkCache::encodeSecTrustRef): Deleted.
(WebKit::NetworkCache::decodeSecTrustRef): Deleted.
(WebKit::NetworkCache::encodeCertificateChain): Deleted.
(WebKit::NetworkCache::decodeCertificateChain): Deleted.
(WebKit::NetworkCache::Coder<WebCore::CertificateInfo>::encode): Deleted.
(WebKit::NetworkCache::Coder<WebCore::CertificateInfo>::decode): Deleted.
* NetworkProcess/cache/NetworkCacheDecoder.cpp: Removed.
* NetworkProcess/cache/NetworkCacheDecoder.h: Removed.
* NetworkProcess/cache/NetworkCacheEncoder.cpp: Removed.
* NetworkProcess/cache/NetworkCacheEncoder.h: Removed.
* NetworkProcess/cache/NetworkCacheEntry.cpp:
(WebKit::NetworkCache::Entry::encodeAsStorageRecord):
(WebKit::NetworkCache::Entry::decodeStorageRecord):
* NetworkProcess/cache/NetworkCacheKey.cpp:
(WebKit::NetworkCache::Key::encode):
(WebKit::NetworkCache::Key::decode):
* NetworkProcess/cache/NetworkCacheKey.h:
* NetworkProcess/cache/NetworkCacheStorage.cpp:
(WebKit::NetworkCache::decodeRecordMetaData):
(WebKit::NetworkCache::encodeRecordMetaData):
* NetworkProcess/cache/NetworkCacheSubresourcesEntry.cpp:
(WebKit::NetworkCache::SubresourceInfo::encode):
(WebKit::NetworkCache::SubresourceInfo::decode):
(WebKit::NetworkCache::SubresourcesEntry::encodeAsStorageRecord):
(WebKit::NetworkCache::SubresourcesEntry::decodeStorageRecord):
* NetworkProcess/cache/NetworkCacheSubresourcesEntry.h:
* UIProcess/API/APIUserContentExtensionStore.cpp:
(API::encodeContentExtensionMetaData):
(API::decodeContentExtensionMetaData):
* WebKit2.xcodeproj/project.pbxproj:

Source/WTF:

Make it possible to use robust serialization of WTF types on the lower levels of the stack.

* WTF.xcodeproj/project.pbxproj:
* wtf/persistence: Added.
* wtf/persistence/Coder.h: Copied from Source/WebKit2/NetworkProcess/cache/NetworkCacheCoder.h.
(WebKit::NetworkCache::Coder::encode): Deleted.
(WebKit::NetworkCache::Coder::decode): Deleted.
* wtf/persistence/Coders.cpp: Copied from Source/WebKit2/NetworkProcess/cache/NetworkCacheCoders.cpp.
(WebKit::NetworkCache::Coder<AtomicString>::encode): Deleted.
(WebKit::NetworkCache::Coder<AtomicString>::decode): Deleted.
(WebKit::NetworkCache::Coder<CString>::encode): Deleted.
(WebKit::NetworkCache::Coder<CString>::decode): Deleted.
(WebKit::NetworkCache::Coder<String>::encode): Deleted.
(WebKit::NetworkCache::decodeStringText): Deleted.
(WebKit::NetworkCache::Coder<String>::decode): Deleted.
(WebKit::NetworkCache::Coder<SHA1::Digest>::encode): Deleted.
(WebKit::NetworkCache::Coder<SHA1::Digest>::decode): Deleted.
(WebKit::NetworkCache::Coder<WebCore::HTTPHeaderMap>::encode): Deleted.
(WebKit::NetworkCache::Coder<WebCore::HTTPHeaderMap>::decode): Deleted.
* wtf/persistence/Coders.h: Copied from Source/WebKit2/NetworkProcess/cache/NetworkCacheCoders.h.
(WebKit::NetworkCache::Coder<std::optional<T>>::encode): Deleted.
(WebKit::NetworkCache::Coder<std::optional<T>>::decode): Deleted.
* wtf/persistence/Decoder.cpp: Copied from Source/WebKit2/NetworkProcess/cache/NetworkCacheDecoder.cpp.
(WebKit::NetworkCache::Decoder::Decoder): Deleted.
(WebKit::NetworkCache::Decoder::~Decoder): Deleted.
(WebKit::NetworkCache::Decoder::bufferIsLargeEnoughToContain): Deleted.
(WebKit::NetworkCache::Decoder::decodeFixedLengthData): Deleted.
(WebKit::NetworkCache::Decoder::decodeNumber): Deleted.
(WebKit::NetworkCache::Decoder::decode): Deleted.
(WebKit::NetworkCache::Decoder::verifyChecksum): Deleted.
* wtf/persistence/Decoder.h: Copied from Source/WebKit2/NetworkProcess/cache/NetworkCacheDecoder.h.
(WebKit::NetworkCache::Decoder::length): Deleted.
(WebKit::NetworkCache::Decoder::currentOffset): Deleted.
(WebKit::NetworkCache::Decoder::decodeEnum): Deleted.
(WebKit::NetworkCache::Decoder::decode): Deleted.
(WebKit::NetworkCache::Decoder::bufferIsLargeEnoughToContain): Deleted.
* wtf/persistence/Encoder.cpp: Copied from Source/WebKit2/NetworkProcess/cache/NetworkCacheEncoder.cpp.
(WebKit::NetworkCache::Encoder::Encoder): Deleted.
(WebKit::NetworkCache::Encoder::~Encoder): Deleted.
(WebKit::NetworkCache::Encoder::grow): Deleted.
(WebKit::NetworkCache::Encoder::updateChecksumForData): Deleted.
(WebKit::NetworkCache::Encoder::encodeFixedLengthData): Deleted.
(WebKit::NetworkCache::Encoder::encodeNumber): Deleted.
(WebKit::NetworkCache::Encoder::encode): Deleted.
(WebKit::NetworkCache::Encoder::encodeChecksum): Deleted.
* wtf/persistence/Encoder.h: Copied from Source/WebKit2/NetworkProcess/cache/NetworkCacheEncoder.h.
(WebKit::NetworkCache::Encoder::encodeEnum): Deleted.
(WebKit::NetworkCache::Encoder::encode): Deleted.
(WebKit::NetworkCache::Encoder::operator<<): Deleted.
(WebKit::NetworkCache::Encoder::buffer): Deleted.
(WebKit::NetworkCache::Encoder::bufferSize): Deleted.
(WebKit::NetworkCache::Encoder::updateChecksumForNumber): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210502 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[GTK] WebProcess from WebKitGtk+ 2.15.2 SIGSEGVs in std::unique_ptr<SoupBuffer, WTF...
carlosgc@webkit.org [Mon, 9 Jan 2017 12:12:14 +0000 (12:12 +0000)]
[GTK] WebProcess from WebKitGtk+ 2.15.2 SIGSEGVs in std::unique_ptr<SoupBuffer, WTF::GPtrDeleter<SoupBuffer> >::get() const () at /usr/include/c++/6/bits/unique_ptr.h:305
https://bugs.webkit.org/show_bug.cgi?id=165848

Reviewed by Michael Catanzaro.

In r208881 several locks were added to ImageDecoder to prevent frameBufferAtIndex() from being called by multiple
threads at the same time, but I forgot isSizeAvailable() also calls frameBufferAtIndex(). However, what we
really need to protect is the GIFImageDecoder, to never allow decoding from more than one thread at the same
time. This patch reverts r208881 and adds a lock to GIFImageDecoder::decode() instead.

* platform/image-decoders/ImageDecoder.cpp:
(WebCore::ImageDecoder::frameIsCompleteAtIndex):
(WebCore::ImageDecoder::frameDurationAtIndex):
(WebCore::ImageDecoder::createFrameImageAtIndex):
* platform/image-decoders/ImageDecoder.h:
* platform/image-decoders/gif/GIFImageDecoder.cpp:
(WebCore::GIFImageDecoder::decode):
* platform/image-decoders/gif/GIFImageDecoder.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210501 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[GTK] Replace WebKitWebSettings with WebKitSettings in documentation
tpopela@redhat.com [Mon, 9 Jan 2017 11:25:49 +0000 (11:25 +0000)]
[GTK] Replace WebKitWebSettings with WebKitSettings in documentation
https://bugs.webkit.org/show_bug.cgi?id=166836

Reviewed by Carlos Garcia Campos.

Don't use WebKitWebSettings (it was a WebKit1 thing) in documentation,
but replace it with WebKitSettings.

* UIProcess/API/gtk/WebKitSettings.cpp:
(webkit_settings_class_init):
* UIProcess/API/gtk/WebKitWebView.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210500 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[OWR] Unskip fast/mediastream/MediaStream-video-element-track-stop.html
commit-queue@webkit.org [Mon, 9 Jan 2017 11:03:35 +0000 (11:03 +0000)]
[OWR] Unskip fast/mediastream/MediaStream-video-element-track-stop.html
https://bugs.webkit.org/show_bug.cgi?id=165316

Patch by Alejandro G. Castro <alex@igalia.com> on 2017-01-09
Reviewed by Philippe Normand.

Source/WebCore:

Fixed the ended support of the video element in the OWR player and
fixed the videoTracks support. Now the OW player properly adds and
removes the audio and video tracks. Added the getSettings support
to the mediastream interface. Solved also the size handling in
some of the enable/muted situations.

Unskipping fast/mediastream/MediaStream-video-element-track-stop.html.

* platform/GStreamer.cmake: Added the new
RealtimeMediaSourceOwr.cpp with the new code handling the
settings.
* platform/graphics/gstreamer/MediaPlayerPrivateGStreamerOwr.cpp:
(WebCore::MediaPlayerPrivateGStreamerOwr::~MediaPlayerPrivateGStreamerOwr):
Release the new video and audio maps that creates a relationship
of the mediastream tracks with the mediatracks of the video element.
(WebCore::MediaPlayerPrivateGStreamerOwr::play): Set ended to
false when we start playing.
(WebCore::MediaPlayerPrivateGStreamerOwr::load): Create the media
tracks to the player so that the videoTracks API returns it, and
add an entry in the map to be able to restore it using the
mediastream track.
(WebCore::MediaPlayerPrivateGStreamerOwr::disableMediaTracks):
Move some part of the stop function to this one in order to use it
in pause method and avoid changing the video selection in that
case.
(WebCore::MediaPlayerPrivateGStreamerOwr::stop): Now we call the
function disableMediaTracks and we also change teh selected
videoTrack.
(WebCore::MediaPlayerPrivateGStreamerOwr::trackEnded): For
videoTrack reset the size and the selected element. In case there
is no media playing we Make sure we set the ended variable to true
and call the timeChange to modify the state of the player.
(WebCore::MediaPlayerPrivateGStreamerOwr::trackEnabledChanged):
Handle properly the situation when the user changed the enabled
value, we disable the media.
(WebCore::MediaPlayerPrivateGStreamerOwr::setSize): Modify the
size of the source.
(WebCore::MediaPlayerPrivateGStreamerOwr::naturalSize): Overrided
to make sure we return a size even when the sample is not ready
for gst.
* platform/graphics/gstreamer/MediaPlayerPrivateGStreamerOwr.h:
Added the ended attribute and the maps.
* platform/mediastream/openwebrtc/MediaEndpointOwr.cpp:
(WebCore::MediaEndpointOwr::createMutedRemoteSource): Use the new
realtime video and audio sources classes, this new classes
implement the settings of each type of media element.
* platform/mediastream/openwebrtc/RealtimeMediaSourceCenterOwr.cpp:
(WebCore::RealtimeMediaSourceCenterOwr::mediaSourcesAvailable):
Use the new audio and video source classes instead of the general
one in order to handle the settings properly.
* platform/mediastream/openwebrtc/RealtimeMediaSourceOwr.cpp: Added
(WebCore::RealtimeMediaSourceOwr::settings): Initialize using the
subclass and return the new currentSettings attribute.
(WebCore::RealtimeMediaSourceOwr::supportedConstraints): Call the
subclass initialization of the supportedSettings.
* platform/mediastream/openwebrtc/RealtimeMediaSourceOwr.h:
(WebCore::RealtimeMediaSourceOwr::initializeSettings): Added to
initialize the settings before returning the value. Implemented in
the subclass.
(WebCore::RealtimeMediaSourceOwr::initializeSupportedConstraints):
Added to initialize the supported settings of the
media. Implemented in the subclass.
(WebCore::RealtimeMediaSourceOwr::settings): Moved to the cpp
file, implemented using the initialize functions of the subclass
* platform/mediastream/openwebrtc/RealtimeVideoSourceOwr.h: Added
this new class to handle the settings of the video elements. The
class initializes the settings and the supportedSettings.
* platform/mediastream/openwebrtc/RealtimeAudioSourceOwr.h: Added
this new class to handle the settings of the audio elements. The
class initializes the settings and the supportedSettings.

LayoutTests:

* platform/gtk/TestExpectations: Unskipped the test and moved the
stop one, now it does not timeout because we correctly handle the
ended value but still fails, we need more features in the mock
classes for owr.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210499 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoInject MarkedSpace size classes for a few more high-volume objects.
akling@apple.com [Mon, 9 Jan 2017 05:35:14 +0000 (05:35 +0000)]
Inject MarkedSpace size classes for a few more high-volume objects.
<https://webkit.org/b/166815>

Reviewed by Darin Adler.

Add the following classes to the list of manually injected size classes:

    - JSString
    - JSFunction
    - PropertyTable
    - Structure

Only Structure actually ends up with a new size class, the others already
can't get any tighter due to the current MarkedBlock::atomSize being 16.
I've put them in anyway to ensure that we have optimally carved-out cells
for them in the future, should they grow.

With this change, Structures get allocated in 128-byte cells instead of
160-byte cells, giving us 25% more Structures per MarkedBlock.

* heap/MarkedSpace.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210498 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoUnreviewed, since this test is slow it should run in fewer configurations. This resol...
fpizlo@apple.com [Mon, 9 Jan 2017 04:12:17 +0000 (04:12 +0000)]
Unreviewed, since this test is slow it should run in fewer configurations. This resolves the
occasional timeouts I was seeing.

* js/script-tests/regress-139548.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210497 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoIntroduce CPU(X86_SSE2) instead of various SSE2 checks
annulen@yandex.ru [Sun, 8 Jan 2017 17:44:57 +0000 (17:44 +0000)]
Introduce CPU(X86_SSE2) instead of various SSE2 checks
https://bugs.webkit.org/show_bug.cgi?id=166808

Reviewed by Michael Catanzaro.

Now copyLCharsFromUCharSource can use SSE2 implementation on non-Darwin
OSes, and all SSE2 code paths are available for MSVC on x86 if /arch:SSE2
or higher is enabled, and for MSVC on x86_64.

Source/WebCore:

No new tests needed.

* platform/audio/SincResampler.cpp:
(WebCore::SincResampler::process):
* platform/audio/VectorMath.cpp:
(WebCore::VectorMath::vsma):
(WebCore::VectorMath::vsmul):
(WebCore::VectorMath::vadd):
(WebCore::VectorMath::vmul):
(WebCore::VectorMath::zvmul):
(WebCore::VectorMath::vsvesq):
(WebCore::VectorMath::vmaxmgv):

Source/WTF:

* wtf/Platform.h:
* wtf/text/ASCIIFastPath.h:
(WTF::copyLCharsFromUCharSource):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210496 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoAvoid triggering rebuilds for minor changes of CSSProperties.json
simon.fraser@apple.com [Sun, 8 Jan 2017 06:02:25 +0000 (06:02 +0000)]
Avoid triggering rebuilds for minor changes of CSSProperties.json
https://bugs.webkit.org/show_bug.cgi?id=166810

Post-review cleanup suggested by Darin Adler.

* css/makeprop.pl:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210495 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoAdd a setting to create editable MiniBrowser windows by default
timothy_horton@apple.com [Sun, 8 Jan 2017 01:22:46 +0000 (01:22 +0000)]
Add a setting to create editable MiniBrowser windows by default
https://bugs.webkit.org/show_bug.cgi?id=166806

Reviewed by Beth Dakin.

* MiniBrowser/mac/AppDelegate.m:
(-[BrowserAppDelegate createBrowserWindowController:]):
(-[BrowserAppDelegate applicationDidFinishLaunching:]):
(-[BrowserAppDelegate _updateNewWindowKeyEquivalents]):
* MiniBrowser/mac/SettingsController.h:
* MiniBrowser/mac/SettingsController.m:
(-[SettingsController _populateMenu]):
(-[SettingsController validateMenuItem:]):
(-[SettingsController toggleCreateEditorByDefault:]):
(-[SettingsController createEditorByDefault]):
Add a setting and update the key equivalents accordingly.
Also make sure to make the right kind of window on launch.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210494 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoAvoid triggering rebuilds for minor changes of CSSProperties.json
simon.fraser@apple.com [Sat, 7 Jan 2017 23:39:32 +0000 (23:39 +0000)]
Avoid triggering rebuilds for minor changes of CSSProperties.json
https://bugs.webkit.org/show_bug.cgi?id=166810

Reviewed by Daniel Bates.

We expect to make lots of metadata-related changes to CSSProperties.json, and
these should not trigger rebuilds.

Have makeprop.pl write to temporary files first, then only replace the generated
files if the contents differ.

* css/makeprop.pl:
(replaceFileIfChanged):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210493 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoRemove PassRefPtr use from "inspector", "platform/sql", and "platform/text"
darin@apple.com [Sat, 7 Jan 2017 22:08:28 +0000 (22:08 +0000)]
Remove PassRefPtr use from "inspector", "platform/sql", and "platform/text"
https://bugs.webkit.org/show_bug.cgi?id=166797

Reviewed by Alex Christensen.

Source/WebCore:

* Modules/webdatabase/Database.cpp:
(WebCore::Database::performOpenAndVerify): Pass a reference instead of a pointer.

* dom/Document.cpp:
(WebCore::canonicalizedTitle): Map backslashes to currency symbols inside the
existing loop through the characters instead of calling a function confusingly
named "displayBuffer" to do it. More straightforward, possibly more efficient,
one less function.
(WebCore::Document::displayStringModifiedByEncoding): Call replace to map
backslashes to currency symbols instead of calling a special displayString
function. There was no real benefit to using that function; the String::replace
function has all the same optimizations, and this was also the only call site.

* editing/SpellChecker.cpp:
(WebCore::SpellChecker::invokeRequest): Pass a reference instead of a pointer.

* inspector/InspectorOverlay.cpp:
(WebCore::InspectorOverlay::overlayPage): Call createEmptyEditorClient instead
of using the class EmptyEditorClient directly.

* loader/EmptyClients.cpp: Moved most of the empty clients here from the header.
Also made lots more things private.
(WebCore::createEmptyEditorClient): Added.
(WebCore::emptyDiagnosticLoggingClient): Added.
* loader/EmptyClients.h: Updated for above changes. Made more things private.

* page/DatabaseProvider.h: Removed unneeded include.
* page/MainFrame.cpp: Ditto.

* page/Page.cpp:
(WebCore::Page::diagnosticLoggingClient): Call emptyDiagnosticLoggingClient
instead of using the class EmptyDiagnosticLoggingClient directly.

* platform/sql/SQLiteDatabase.cpp:
(WebCore::SQLiteDatabase::setAuthorizer): Take a reference instead of a
PassRefPtr.
* platform/sql/SQLiteDatabase.h: Updated for above.

* platform/text/BidiContext.cpp:
(WebCore::BidiContext::BidiContext): Moved this function here from the header.
(WebCore::copyContextAndRebaselineLevel): Return a Ref instead of a PassRefPtr,
and take a reference rather than a raw pointer.
(WebCore::BidiContext::copyStackRemovingUnicodeEmbeddingContexts): Return a
Ref instead of a PassRefPtr.
* platform/text/BidiContext.h: Updated for the above changes.

* platform/text/BidiResolver.h: Simplify initialization, use more references,
auto, modern for loops, refer to derived class as that rather than the non-C++
term subclass.

* platform/text/TextCheckerClient.h: Changd argument from PassRefPtr to a
reference.

* platform/text/TextEncoding.h:
(WebCore::TextEncoding::displayString): Deleted.
(WebCore::TextEncoding::displayBuffer): Deleted.

* rendering/RootInlineBox.cpp: Style tweaks and use a modern for loop.

* svg/graphics/SVGImage.cpp:
(WebCore::SVGImage::dataChanged): Call createEmptyEditorClient instead
of using the class EmptyEditorClient directly.

Source/WebKit/mac:

* WebCoreSupport/WebEditorClient.h: Updated to take a reference.
* WebCoreSupport/WebEditorClient.mm:
(WebEditorClient::requestCheckingOfString): Take a reference
instead of a PassRefPtr.

Source/WebKit/win:

* WebCoreSupport/WebEditorClient.h: Update argument to take a reference
instead of a PassRefPtr.

Source/WebKit2:

* WebProcess/WebCoreSupport/WebEditorClient.cpp:
(WebKit::WebEditorClient::requestCheckingOfString): Take a reference
instead of a PassRefPtr.
* WebProcess/WebCoreSupport/WebEditorClient.h: Updated for above change.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210492 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[QuickLook] Update quicklook test expectations
aestes@apple.com [Sat, 7 Jan 2017 21:29:57 +0000 (21:29 +0000)]
[QuickLook] Update quicklook test expectations
https://bugs.webkit.org/show_bug.cgi?id=166798
<rdar://problem/26477735>

Reviewed by Alex Christensen.

Source/WebCore:

QuickLook uses UUIDs to generate x-apple-ql-id: URLs, so they will differ each time a
document is previewed. Since these URLs are included in QuickLook layout test results,
WebArchiveDumpSupport needs to replace them with predictable values when dumping a web
archive.

* testing/cocoa/WebArchiveDumpSupport.mm:
(WebCoreTestSupport::quickLookURLReplacements): Created a static Vector of CFStringRef pairs
to store x-apple-ql-id: URLs and their replacement values.
(WebCoreTestSupport::normalizeWebResourceURL): If the URL starts with “x-apple-ql-id://”,
replace the text between the scheme and the file extension with “resource”. Store the
original URL along with its replacement in quickLookURLReplacements().
(WebCoreTestSupport::createXMLStringFromWebArchiveData): Replaced all QuickLook URLs in the
final XML text with their replacement URLs.

LayoutTests:

For documents that generate HTML previews, updated the expected results to be web archives.
Web archives capture the same information that the expected HTML files did, but allow for
easy rebasing with --reset-results.

For documents that generate PDF previews, a webarchive expected result is inappropriate
since the generated PDF will differ between runs due to changing metadata (e.g. creation
date). These tests were kept as reftests but with updated expected results.

* platform/ios-simulator/TestExpectations: Unskipped quicklook/.
* quicklook/excel-expected.html: Removed.
* quicklook/excel-expected.txt: Added.
* quicklook/excel-legacy-expected.html: Removed.
* quicklook/excel-legacy-expected.txt: Added.
* quicklook/excel-legacy.html: Updated to dump as webarchive.
* quicklook/excel.html: Ditto.
* quicklook/keynote-09-expected.html: Removed.
* quicklook/keynote-09-expected.txt: Added.
* quicklook/keynote-09.html: Updated to dump as webarchive.
* quicklook/numbers-09-expected.html: Removed.
* quicklook/numbers-09-expected.txt: Added.
* quicklook/numbers-09.html: Updated to dump as webarchive.
* quicklook/pages-09-expected.html: Removed.
* quicklook/pages-09-expected.txt: Added.
* quicklook/pages-09.html: Updated to dump as webarchive.
* quicklook/powerpoint-expected.html: Removed.
* quicklook/powerpoint-expected.txt: Added.
* quicklook/powerpoint-legacy-expected.html: Removed.
* quicklook/powerpoint-legacy-expected.txt: Added.
* quicklook/powerpoint-legacy.html: Updated to dump as webarchive.
* quicklook/powerpoint.html: Updated to dump as webarchive.
* quicklook/resources/excel-expected.html: Removed.
* quicklook/resources/excel-legacy-expected.html: Removed.
* quicklook/resources/keynote-09-expected/index.css: Removed.
* quicklook/resources/keynote-09-expected/index.html: Removed.
* quicklook/resources/keynote-09-expected/index.js: Removed.
* quicklook/resources/keynote-expected.pdf: Updated.
* quicklook/resources/numbers-09-expected/canvas.js: Removed.
* quicklook/resources/numbers-09-expected/index.html: Removed.
* quicklook/resources/numbers-09-expected/navigation.css: Removed.
* quicklook/resources/numbers-09-expected/navigation.html: Removed.
* quicklook/resources/numbers-09-expected/sheet_1.html: Removed.
* quicklook/resources/numbers-expected.pdf: Updated.
* quicklook/resources/pages-09-expected/index.css: Removed.
* quicklook/resources/pages-09-expected/index.html: Removed.
* quicklook/resources/pages-expected.pdf: Updated.
* quicklook/resources/powerpoint-expected.html: Removed.
* quicklook/resources/powerpoint-legacy-expected.html: Removed.
* quicklook/resources/webkit-icon.pdf: Removed.
* quicklook/resources/webkit-icon.png: Removed.
* quicklook/resources/webkit-icon.tiff: Removed.
* quicklook/resources/word-expected.html: Removed.
* quicklook/resources/word-legacy-expected.html: Removed.
* quicklook/word-expected.html: Removed.
* quicklook/word-expected.txt: Added.
* quicklook/word-legacy-expected.html: Removed.
* quicklook/word-legacy-expected.txt: Added.
* quicklook/word-legacy.html: Updated to dump as webarchive.
* quicklook/word.html: Updated to dump as webarchive.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210491 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoFix the 32-bit build.
timothy_horton@apple.com [Sat, 7 Jan 2017 21:23:39 +0000 (21:23 +0000)]
Fix the 32-bit build.

* MiniBrowser/mac/BrowserWindowController.h:
* MiniBrowser/mac/BrowserWindowController.m:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210490 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoFix the 32-bit build.
timothy_horton@apple.com [Sat, 7 Jan 2017 21:18:13 +0000 (21:18 +0000)]
Fix the 32-bit build.

* MiniBrowser/mac/BrowserWindowController.m:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210489 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoMake it possible to drop HTML files on the MiniBrowser icon
timothy_horton@apple.com [Sat, 7 Jan 2017 21:15:55 +0000 (21:15 +0000)]
Make it possible to drop HTML files on the MiniBrowser icon
https://bugs.webkit.org/show_bug.cgi?id=166802

Reviewed by Simon Fraser.

* MiniBrowser/mac/AppDelegate.m:
(-[BrowserAppDelegate application:openFile:]):
* MiniBrowser/mac/Info.plist:
If you're using an editable window, the usual trick of dropping on the
window doesn't work anymore (it makes an <attachment> instead).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210488 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoMiniBrowser window gets smaller each time it is instantiated
timothy_horton@apple.com [Sat, 7 Jan 2017 21:13:10 +0000 (21:13 +0000)]
MiniBrowser window gets smaller each time it is instantiated
https://bugs.webkit.org/show_bug.cgi?id=166807

Reviewed by Andy Estes.

* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController windowDidLoad]):
* MiniBrowser/mac/BrowserWindow.xib:
Turn on full-window content view support with IB, instead of waiting
until after the views have been loaded. This seems to resolve the
longstanding ever-shrinking-window problem.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210487 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoAdd indent/outdent items to MiniBrowser's new format menu
timothy_horton@apple.com [Sat, 7 Jan 2017 21:12:26 +0000 (21:12 +0000)]
Add indent/outdent items to MiniBrowser's new format menu
https://bugs.webkit.org/show_bug.cgi?id=166805

Reviewed by Dan Bernstein.

* MiniBrowser/mac/MainMenu.xib:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210486 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoAdd a stock Format menu to MiniBrowser
timothy_horton@apple.com [Sat, 7 Jan 2017 21:11:53 +0000 (21:11 +0000)]
Add a stock Format menu to MiniBrowser
https://bugs.webkit.org/show_bug.cgi?id=166804

Reviewed by Andy Estes.

* MiniBrowser/mac/MainMenu.xib:
Many of the features automatically work with editable web views.
However, this seems to expose a bug where none of the font-related features
(Cmd-B, for example) work until the first time the font panel is shown,
per process, which will need to be investigated separately.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210485 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoRemove unused applicationWillTerminate methods from MiniBrowser
timothy_horton@apple.com [Sat, 7 Jan 2017 21:11:02 +0000 (21:11 +0000)]
Remove unused applicationWillTerminate methods from MiniBrowser
https://bugs.webkit.org/show_bug.cgi?id=166803

Reviewed by Andy Estes.

* MiniBrowser/mac/AppDelegate.m:
(-[BrowserAppDelegate applicationWillTerminate:]): Deleted.
* MiniBrowser/mac/BrowserWindowController.h:
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController applicationTerminating]): Deleted.
* MiniBrowser/mac/WK1BrowserWindowController.m:
(-[WK1BrowserWindowController applicationTerminating]): Deleted.
* MiniBrowser/mac/WK2BrowserWindowController.m:
(-[WK2BrowserWindowController applicationTerminating]): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210484 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoAdd the ability to create empty contentEditable windows in MiniBrowser
timothy_horton@apple.com [Sat, 7 Jan 2017 21:10:21 +0000 (21:10 +0000)]
Add the ability to create empty contentEditable windows in MiniBrowser
https://bugs.webkit.org/show_bug.cgi?id=166801

Reviewed by Wenson Hsieh.

* MiniBrowser/mac/AppDelegate.h:
* MiniBrowser/mac/AppDelegate.m:
(-[BrowserAppDelegate createBrowserWindowController:]):
(-[BrowserAppDelegate newWindow:]):
(-[BrowserAppDelegate newEditorWindow:]):
(-[BrowserAppDelegate openDocument:]):
(-[BrowserAppDelegate _updateNewWindowKeyEquivalents]):
* MiniBrowser/mac/BrowserWindowController.h:
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController loadHTMLString:]):
(-[BrowserWindowController toggleEditable:]):
* MiniBrowser/mac/MainMenu.xib:
* MiniBrowser/mac/WK1BrowserWindowController.m:
(-[WK1BrowserWindowController awakeFromNib]):
(-[WK1BrowserWindowController loadHTMLString:]):
(-[WK1BrowserWindowController validateMenuItem:]):
(-[WK1BrowserWindowController setEditable:]):
(-[WK1BrowserWindowController updateTitle:]):
* MiniBrowser/mac/WK2BrowserWindowController.m:
(-[WK2BrowserWindowController awakeFromNib]):
(-[WK2BrowserWindowController validateMenuItem:]):
(-[WK2BrowserWindowController setEditable:]):
(-[WK2BrowserWindowController updateTitle:]):
(-[WK2BrowserWindowController loadHTMLString:]):
Merge more functionality from WebEditingTester.
Add two menu items for creating empty editable WebKit1/2 windows.
Add a menu item to the Edit menu to toggle editability on any window.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210483 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoImplement find-in-page in MiniBrowser
timothy_horton@apple.com [Sat, 7 Jan 2017 21:09:41 +0000 (21:09 +0000)]
Implement find-in-page in MiniBrowser
https://bugs.webkit.org/show_bug.cgi?id=166799

Reviewed by Andy Estes.

* MiniBrowser/mac/BrowserWindow.xib:
* MiniBrowser/mac/BrowserWindowController.h:
* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController find:]): Deleted.
* MiniBrowser/mac/MainMenu.xib:
* MiniBrowser/mac/WK1BrowserWindowController.m:
(-[WK1BrowserWindowController toggleShrinkToFit:]):
(-[WK1BrowserWindowController find:]): Deleted.
* MiniBrowser/mac/WK2BrowserWindowController.m:
(-[WK2BrowserWindowController awakeFromNib]):
(-[WK2BrowserWindowController dealloc]):
(-[WK2BrowserWindowController validateMenuItem:]):
(-[WK2BrowserWindowController performTextFinderAction:]):
(-[WK2BrowserWindowController findBarView]):
(-[WK2BrowserWindowController setFindBarView:]):
(-[WK2BrowserWindowController isFindBarVisible]):
(-[WK2BrowserWindowController setFindBarVisible:]):
(-[WK2BrowserWindowController contentView]):
(-[WK2BrowserWindowController findBarViewDidChangeHeight]):
(-[WK2BrowserWindowController performFindPanelAction:]): Deleted.
(-[WK2BrowserWindowController find:]): Deleted.
Implement find-in-page using WKWebView's conformance to the NSTextFinderClient
protocol. At first glance, it's pretty broken, but vaguely works.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210482 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoShare the implementation of MiniBrowser view hiding functionality
timothy_horton@apple.com [Sat, 7 Jan 2017 21:08:43 +0000 (21:08 +0000)]
Share the implementation of MiniBrowser view hiding functionality
https://bugs.webkit.org/show_bug.cgi?id=166800

Reviewed by Andy Estes.

* MiniBrowser/mac/BrowserWindowController.m:
(-[BrowserWindowController showHideWebView:]):
(-[BrowserWindowController removeReinsertWebView:]):
* MiniBrowser/mac/WK1BrowserWindowController.m:
(-[WK1BrowserWindowController showHideWebView:]): Deleted.
(-[WK1BrowserWindowController removeReinsertWebView:]): Deleted.
* MiniBrowser/mac/WK2BrowserWindowController.m:
(-[WK2BrowserWindowController showHideWebView:]): Deleted.
(-[WK2BrowserWindowController removeReinsertWebView:]): Deleted.
Move the implementation of these two things to the shared superclass,
since they don't depend on WebKit-specific things.
Also, there was no reason to retain the (WK)WebView while it was unparented,
as the BrowserWindowController maintains a reference the entire time.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210481 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years ago[QuickLook] Consolidate usage of QuickLookHandle into SubresourceLoader
aestes@apple.com [Sat, 7 Jan 2017 21:00:30 +0000 (21:00 +0000)]
[QuickLook] Consolidate usage of QuickLookHandle into SubresourceLoader
https://bugs.webkit.org/show_bug.cgi?id=166713

Reviewed by Alex Christensen.

Source/WebCore:

QuickLook conversion was originally implemented by intercepting document loads at the
ResourceHandle level, with separate paths for NSURLConnection and CFURLConnection handles.
When QuickLook was ported to WebKit2, a third path was added in WebResourceLoader.

This change removes these three separate paths and implements a single conversion path in
SubresourceLoader, where QuickLook can intercept document loads regardless of the networking
API being used.

No change in behavior. Covered by existing tests.

* loader/ResourceLoader.h: Removed override from didCreateQuickLookHandle() since this
function is no longer declared in ResourceHandleClient.
* loader/SubresourceLoader.cpp:
(WebCore::SubresourceLoader::shouldCreateQuickLookHandleForResponse): Added a helper
function to check if a QuickLookHandle should be created for a response.
(WebCore::SubresourceLoader::didReceiveResponse): Created a QuickLookHandle if necessary
and stored it in the DocumentLoader.
(WebCore::SubresourceLoader::didReceiveData): If there is a QuickLookHandle, call
QuickLookHandle::didReceiveBuffer(), and return early if QuickLook is converting.
(WebCore::SubresourceLoader::didReceiveBuffer): Ditto for QuickLookHandle::didReceiveBuffer().
(WebCore::SubresourceLoader::didFinishLoading): Ditto for QuickLookHandle::didFinishLoading().
(WebCore::SubresourceLoader::didFail): If there is a QuickLookHandle, call QuickLookHandle::didFail().
* loader/SubresourceLoader.h: Declared shouldCreateQuickLookHandleForResponse().
* loader/ios/QuickLook.h: Removed declarations used by the ResourceHandles and declared new
functions used by SubresourceLoader.
* loader/ios/QuickLook.mm:
(WebCore::QuickLookHandle::QuickLookHandle): Removed the NSURLConnection * parameter, since
we’d now always pass nil.
(WebCore::QuickLookHandle::create): Removed create() functions used by ResourceHandles.
(WebCore::QuickLookHandle::didReceiveData): Added. Copies the data into an NSData, wraps it
in an NSArray, and passes it to didReceiveDataArray().
(WebCore::QuickLookHandle::didReceiveBuffer): Added. Creates a NSArray of NSData from the
SharedBuffer and passes it to didReceiveDataArray().
(-[WebQuickLookHandleAsDelegate initWithConnectionDelegate:]): Deleted.
(-[WebQuickLookHandleAsDelegate connection:didReceiveDataArray:]): Deleted.
(-[WebQuickLookHandleAsDelegate connection:didReceiveData:lengthReceived:]): Deleted.
(-[WebQuickLookHandleAsDelegate connectionDidFinishLoading:]): Deleted.
(-[WebQuickLookHandleAsDelegate connection:didFailWithError:]): Deleted.
(-[WebQuickLookHandleAsDelegate detachHandle]): Deleted.
(WebCore::QuickLookHandle::cfResponse): Deleted.
(WebCore::QuickLookHandle::didReceiveDataArray): Passed dataArray to m_converter and m_client.
* platform/SharedBuffer.h: Const-qualified createNSDataArray().
* platform/cocoa/SharedBufferCocoa.mm:
(WebCore::SharedBuffer::createNSDataArray): Ditto.
* platform/network/ResourceHandle.cpp: Stopped including QuickLook.h.
* platform/network/ResourceHandle.h: Removed m_quickLook.
(WebCore::ResourceHandle::quickLookHandle): Deleted.
* platform/network/ResourceHandleClient.h:
(WebCore::ResourceHandleClient::didCreateQuickLookHandle): Deleted.
* platform/network/cf/SynchronousResourceHandleCFURLConnectionDelegate.cpp:
(WebCore::SynchronousResourceHandleCFURLConnectionDelegate::didReceiveResponse): Removed QuickLook code.
(WebCore::SynchronousResourceHandleCFURLConnectionDelegate::didReceiveData): Ditto.
(WebCore::SynchronousResourceHandleCFURLConnectionDelegate::didFinishLoading): Ditto.
(WebCore::SynchronousResourceHandleCFURLConnectionDelegate::didFail): Ditto.
(WebCore::SynchronousResourceHandleCFURLConnectionDelegate::didReceiveDataArray): Ditto.
* platform/network/mac/ResourceHandleMac.mm:
(WebCore::ResourceHandle::setQuickLookHandle): Deleted.
* platform/network/mac/WebCoreResourceHandleAsDelegate.mm:
(-[WebCoreResourceHandleAsDelegate connection:didReceiveResponse:]): Removed QuickLook code.
(-[WebCoreResourceHandleAsDelegate connection:didReceiveDataArray:]): Ditto.
(-[WebCoreResourceHandleAsDelegate connection:didReceiveData:lengthReceived:]): Ditto.
(-[WebCoreResourceHandleAsDelegate connectionDidFinishLoading:]): Ditto.
(-[WebCoreResourceHandleAsDelegate connection:didFailWithError:]): Ditto.

Source/WebKit2:

* WebProcess/Network/WebResourceLoader.cpp:
(WebKit::WebResourceLoader::didReceiveResponse): Removed QuickLook code.
(WebKit::WebResourceLoader::didReceiveData): Ditto.
(WebKit::WebResourceLoader::didFinishResourceLoad): Ditto.
(WebKit::WebResourceLoader::didFailResourceLoad): Ditto.
(WebKit::WebResourceLoader::didReceiveResource): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210480 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoAnother attempt to fix the Mac CMake build after r210467.
aestes@apple.com [Sat, 7 Jan 2017 18:51:46 +0000 (18:51 +0000)]
Another attempt to fix the Mac CMake build after r210467.

* DumpRenderTree/PlatformMac.cmake:
* WebKitTestRunner/PlatformMac.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210479 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2 years agoAttempt to fix the Mac CMake build after r210467.
aestes@apple.com [Sat, 7 Jan 2017 18:30:55 +0000 (18:30 +0000)]
Attempt to fix the Mac CMake build after r210467.

* testing/cocoa/WebArchiveDumpSupport.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@210478 268f45cc-cd09-0410-ab3c-d52691b4dbfc