⚠ Archived content — this site is no longer maintained.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

Timeline



Aug 3, 2018:

11:35 PM Changeset in webkit [234577] by rniwa@webkit.org
  • 7 edits in trunk

innerHTML should not synchronously create a custom element
​https://bugs.webkit.org/show_bug.cgi?id=188327
<rdar://problem/42923114>

Reviewed by Daniel Bates.

LayoutTests/imported/w3c:

Rebaselined the test now that all test cases are passing.

  • web-platform-tests/custom-elements/connected-callbacks-html-fragment-parsing-expected.txt:

Source/WebCore:

Fixed the bug that the fragment parsing algorithm was synchronously constructing a custom element instead of
enqueuing an element to upgrade.

The fragment parsing algorithm creates an element for a token with *will execute script* flag set to false:
​https://html.spec.whatwg.org/multipage/parsing.html#create-an-element-for-the-token
which results in creating an element with synchronous custom elements flag *not* set:
​https://dom.spec.whatwg.org/#concept-create-element

When synchronous custom elements flag is false, we're supposed to create an element and enqueue a custom element
upgrade reaction. createHTMLElementOrFindCustomElementInterface was missing this last logic.

Also fixed a bug that Element::enqueueToUpgrade would hit a debug assertion when a custom element which has been
enqueued to upgrade is enqueued to upgrade for the second time. In this case, we need to put the element into the
current element queue (​https://html.spec.whatwg.org/multipage/custom-elements.html#current-element-queue) again.

While the specification simply enqueues another upgrade reaction and bails out immediately in the first step of
the upgrade, WebKit's implementation simply avoids this redundancy in the first place:
​https://html.spec.whatwg.org/multipage/custom-elements.html#concept-upgrade-an-element

Existing tests such as imported/w3c/web-platform-tests/custom-elements/reactions/Document.html exercises this
code path after the fragment parsing algorithm fix.

Tests: imported/w3c/web-platform-tests/custom-elements/connected-callbacks-html-fragment-parsing.html

  • dom/CustomElementReactionQueue.cpp:

(WebCore::CustomElementReactionQueueItem::type const): Added for an assertion.
(WebCore::CustomElementReactionQueue::enqueueElementUpgrade): Enqueue this element to the current element queue
by calling ensureCurrentQueue and avoid inserting a redundant upgrade reaction.

  • dom/CustomElementReactionQueue.h:
  • dom/Element.cpp:

(WebCore::Element::enqueueToUpgrade): Handle the case when a custom element is enqueued to upgrade for the second
time while it had been waiting in some element queue. In this case, the reaction queue for this element has
already been created and we simply need to put this element back into the current element queue (i.e. this element
now belongs to both element queues).

  • html/parser/HTMLConstructionSite.cpp:

(WebCore::findCustomElementInterface): Extracted out of createHTMLElementOrFindCustomElementInterface.
(WebCore::HTMLConstructionSite::createHTMLElementOrFindCustomElementInterface): Fixed the bug that the HTML parser
was synchronously constructing a custom element even for the fragment parsing algorithm.

11:27 PM Changeset in webkit [234576] by sbarati@apple.com
  • 3 edits
    1 add in trunk/Source/JavaScriptCore

Give the jsc shell the JIT entitlement
​https://bugs.webkit.org/show_bug.cgi?id=188324
<rdar://problem/42885806>

Reviewed by Dan Bernstein.

This should help us in ensuring the system jsc is able to JIT.

  • Configurations/JSC.xcconfig:
  • JavaScriptCore.xcodeproj/project.pbxproj:
  • allow-jit-macOS.entitlements: Added.
7:04 PM Changeset in webkit [234575] by bshafiei@apple.com
  • 1 copy in tags/Safari-606.1.36

Tag Safari-606.1.36.

7:03 PM Changeset in webkit [234574] by bshafiei@apple.com
  • 1 delete in tags/Safari-606.1.36

Delete tag

7:01 PM Changeset in webkit [234573] by bshafiei@apple.com
  • 2 edits in branches/safari-606-branch/Tools

Cherry-pick r234547. rdar://problem/42902608

Unreviewed gardening; r234534 changed the policy for what audio session is set when a
media element is paused. Updating the test to reflect the new policy.

  • TestWebKitAPI/Tests/WebKitLegacy/ios/AudioSessionCategoryIOS.mm: (TestWebKitAPI::TEST):

git-svn-id: ​https://svn.webkit.org/repository/webkit/trunk@234547 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7:01 PM Changeset in webkit [234572] by bshafiei@apple.com
  • 13 edits
    3 adds in branches/safari-606-branch

Cherry-pick r234534. rdar://problem/42902608

Control center controls disappear when pausing, locking device.
​https://bugs.webkit.org/show_bug.cgi?id=188282
<rdar://problem/42497809>

Reviewed by Eric Carlson.

Source/WebCore:

Test: platform/mac/media/audio-session-category-video-paused.html

MediaRemote will set any app whose audio session category changes from MediaPlayback to
Ambient as not eligable for being the "now playing" app. Previously, due to the ordering of
events, we never moved from MediaPlayback to Ambient when pausing <video>, even though that
was the intention. Now that that bug is fixed, it exposed this new issue with MediaRemote.

To solve the new issue, make our audio session category policy more explicit: once we are in
MediaPlayback category, we will remain so as long as the media element in queston stays
loaded, has played, and is never interrupted by the system.

Make MediaSessionManagerCocoa a proper subclass of PlatformMediaSessionManager, and make
MediaSessionManageriOS and -Mac subclasses of the -Cocoa class. Add a new
m_hasPlayedSinceLastInterruption ivar to PlatformMediaSession, set when the state() changes
to Playing, and cleared during an interruption. Check this flag when deciding what audio
sessino category to set in MediaSessionManagerCocoa.

  • WebCore.xcodeproj/project.pbxproj:
  • platform/audio/PlatformMediaSession.cpp: (WebCore::PlatformMediaSession::setState):
  • platform/audio/PlatformMediaSession.h: (WebCore::PlatformMediaSession::hasPlayedSinceLastInterruption const): (WebCore::PlatformMediaSession::clearHasPlayedSinceLastInterruption):
  • platform/audio/PlatformMediaSessionManager.cpp: (WebCore::PlatformMediaSessionManager::updateSessionState): Deleted.
  • platform/audio/PlatformMediaSessionManager.h: (WebCore::PlatformMediaSessionManager::audioHardwareListener): (WebCore::PlatformMediaSessionManager::updateSessionState):
  • platform/audio/cocoa/MediaSessionManagerCocoa.cpp: (MediaSessionManagerCocoa::updateSessionState): (MediaSessionManagerCocoa::beginInterruption): (PlatformMediaSessionManager::updateSessionState): Deleted.
  • platform/audio/cocoa/MediaSessionManagerCocoa.h: Added.
  • platform/audio/ios/MediaSessionManagerIOS.h:
  • platform/audio/ios/MediaSessionManagerIOS.mm: (WebCore::MediaSessionManageriOS::MediaSessionManageriOS):
  • platform/audio/mac/MediaSessionManagerMac.h:
  • platform/audio/mac/MediaSessionManagerMac.mm: (WebCore::MediaSessionManagerMac::MediaSessionManagerMac):

LayoutTests:

  • platform/mac/media/audio-session-category-video-paused-expected.txt: Added.
  • platform/mac/media/audio-session-category-video-paused.html: Added.
  • platform/ios/TestExpectations:

git-svn-id: ​https://svn.webkit.org/repository/webkit/trunk@234534 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6:33 PM Changeset in webkit [234571] by bshafiei@apple.com
  • 1 copy in tags/Safari-606.1.36

Tag Safari-606.1.36.

6:24 PM Changeset in webkit [234570] by bshafiei@apple.com
  • 1 delete in tags/Safari-606.1.36

Delete tag

6:09 PM Changeset in webkit [234569] by commit-queue@webkit.org
  • 23 edits
    5 adds in trunk

We should cache the compiled sandbox profile in a data vault
​https://bugs.webkit.org/show_bug.cgi?id=184991

Patch by Ben Richards <​benton_richards@apple.com> on 2018-08-03
Reviewed by Ryosuke Niwa.

Source/WebCore:

Added functionality to FileHandle so that it can lock a file while open.
Added a function to FileSystem to delete non empty directories.

  • platform/FileHandle.cpp:

(WebCore::FileHandle::FileHandle):
(WebCore::FileHandle::open):
(WebCore::FileHandle::close):

  • platform/FileHandle.h:
  • platform/FileSystem.h:
  • platform/cocoa/FileSystemCocoa.mm:

(WebCore::FileSystem::deleteNonEmptyDirectory):

Source/WebKit:

This patch changes a few things (note: data vaults and sandbox entitlements are only used in internal builds):
(1) Instead of compiling a sandbox every time a process is launched, processes now look for a cached sandbox

in a process specific data vault on macOS platforms. (ChildProcessMac.mm)

(2) If a valid cached sandbox is not found, a process will create the data vault (or ensure that it exists),

compile a sandbox, and cache it.

(3) In order to create process specific data vaults, each process now has their own <process name>-OSX-sandbox.entitlements

file which contains an entitlement with a process specific "storage class" which ensures that each process
can only ever access its own data vault. (See the article on confluence "Data Vaults and Restricted Files" for more info)

(4) The sandbox entitlements file for the Network, WebContent and Plugin services are loaded dynamically

through Scripts/<process name>-process-entitlements.sh which is triggered in a new build phase for each service.
The Storage process sandbox entitlements are loaded directly in Configurations/StorageService.xcconfig.
The reason that the sandbox entitlements are applied dynamically is so that these sandbox entitlements
are only applied when WK_USE_RESTRICTED_ENTITLEMENTS is YES. This means that open source builds will still work.

  • Configurations/Network-OSX-sandbox.entitlements: Added.
  • Configurations/Plugin-OSX-sandbox.entitlements: Added.
  • Configurations/Storage-OSX-sandbox.entitlements: Added.
  • Configurations/StorageService.xcconfig:
  • Configurations/WebContent-OSX-sandbox.entitlements: Added.
  • Configurations/WebKit.xcconfig:
  • NetworkProcess/NetworkProcess.h:
  • PluginProcess/PluginProcess.h:
  • Scripts/process-network-entitlements.sh: Added.
  • Scripts/process-plugin-entitlements.sh: Added.
  • Scripts/process-webcontent-entitlements.sh:
  • Shared/ChildProcess.h:
  • Shared/EntryPointUtilities/mac/XPCService/XPCServiceEntryPoint.h:

(WebKit::XPCServiceInitializer):

  • Shared/SandboxInitializationParameters.h:

(WebKit::SandboxInitializationParameters::setOverrideSandboxProfilePath):
(WebKit::SandboxInitializationParameters::overrideSandboxProfilePath const):
(WebKit::SandboxInitializationParameters::setSandboxProfile):
(WebKit::SandboxInitializationParameters::sandboxProfile const):
(): Deleted.

  • Shared/mac/ChildProcessMac.mm:

(WebKit::SandboxProfileDeleter::operator()):
(WebKit::SandboxParametersDeleter::operator()):
(WebKit::SandboxInfo::SandboxInfo):
(WebKit::fileContents):
(WebKit::processStorageClass):
(WebKit::setAndSerializeSandboxParameters):
(WebKit::getUserCacheDirectory):
(WebKit::sandboxDataVaultParentDirectory):
(WebKit::sandboxDirectory):
(WebKit::sandboxFilePath):
(WebKit::ensureSandboxCacheDirectory):
(WebKit::writeSandboxDataToCacheFile):
(WebKit::compileAndCacheSandboxProfile):
(WebKit::tryApplyCachedSandbox):
(WebKit::webKit2Bundle):
(WebKit::sandboxProfilePath):
(WebKit::compileAndApplySandboxSlowCase):
(WebKit::applySandbox):
(WebKit::initializeSandboxParameters):
(WebKit::ChildProcess::initializeSandbox):

  • Shared/mac/SandboxInitialiationParametersMac.mm:

(WebKit::SandboxInitializationParameters::SandboxInitializationParameters):

  • StorageProcess/StorageProcess.h:
  • WebKit.xcodeproj/project.pbxproj:
  • WebProcess/WebProcess.h:

Source/WTF:

Added trace points for sandbox initialization and exposed functions needed for sandbox caching

  • wtf/SystemTracing.h:
  • wtf/spi/darwin/SandboxSPI.h:

Tools:

Added trace points for sandbox initialization

  • Tracing/SystemTracePoints.plist:
6:06 PM Changeset in webkit [234568] by bshafiei@apple.com
  • 13 edits
    3 deletes in branches/safari-606-branch

Revert r234534. rdar://problem/42902608

6:06 PM Changeset in webkit [234567] by bshafiei@apple.com
  • 2 edits in branches/safari-606-branch/Tools

Revert r234547. rdar://problem/42902608

5:11 PM Changeset in webkit [234566] by commit-queue@webkit.org
  • 24 edits in trunk

WebGL 2 conformance: vertex_arrays/vertex_array_object.html
​https://bugs.webkit.org/show_bug.cgi?id=188291
<rdar://problem/42792709>

Source/WebCore:

Patch by Justin Fan <Justin Fan> on 2018-08-03
Reviewed by Dean Jackson.

Get ​https://www.khronos.org/registry/webgl/conformance-suites/2.0.0/conformance2/vertex_arrays/vertex-array-object.html
working on Mac OS and iOS, and check in the new test expectations.

Existing test: webgl/2.0.0/conformance2/vertex_arrays/vertex-array-object.html

  • html/canvas/WebGL2RenderingContext.cpp:

(WebCore::WebGL2RenderingContext::initializeVertexArrayObjects):
(WebCore::WebGL2RenderingContext::deleteVertexArray):

  • html/canvas/WebGLRenderingContextBase.cpp:

(WebCore::WebGLRenderingContextBase::deleteObject):

  • html/canvas/WebGLVertexArrayObject.cpp:

(WebCore::WebGLVertexArrayObject::WebGLVertexArrayObject):

  • platform/graphics/opengl/GraphicsContext3DOpenGLCommon.cpp:

(WebCore::GraphicsContext3D::checkVaryingsPacking const):
(WebCore::GraphicsContext3D::createVertexArray):
(WebCore::GraphicsContext3D::deleteVertexArray):
(WebCore::GraphicsContext3D::isVertexArray):
(WebCore::GraphicsContext3D::bindVertexArray):

LayoutTests:

Get ​https://www.khronos.org/registry/webgl/conformance-suites/2.0.0/conformance2/vertex_arrays/vertex-array-object.html
working on Mac OS and iOS, and check in the new test expectations.

Patch by Justin Fan <Justin Fan> on 2018-08-03
Reviewed by Dean Jackson.

The following FAILed before this patch but are now expecting PASSes.

  • webgl/2.0.0/conformance2/glsl3/array-as-return-value-expected.txt:
  • webgl/2.0.0/conformance2/glsl3/array-assign-constructor-expected.txt:
  • webgl/2.0.0/conformance2/glsl3/array-assign-expected.txt:
  • webgl/2.0.0/conformance2/glsl3/array-complex-indexing-expected.txt:
  • webgl/2.0.0/conformance2/glsl3/array-element-increment-expected.txt:
  • webgl/2.0.0/conformance2/glsl3/array-equality-expected.txt:
  • webgl/2.0.0/conformance2/glsl3/array-in-complex-expression-expected.txt:
  • webgl/2.0.0/conformance2/glsl3/bool-type-cast-bug-uint-ivec-uvec-expected.txt:
  • webgl/2.0.0/conformance2/glsl3/compare-structs-containing-arrays-expected.txt:
  • webgl/2.0.0/conformance2/glsl3/const-array-init-expected.txt:
  • webgl/2.0.0/conformance2/glsl3/frag-depth-expected.txt:
  • webgl/2.0.0/conformance2/glsl3/loops-with-side-effects-expected.txt:
  • webgl/2.0.0/conformance2/glsl3/no-attribute-vertex-shader-expected.txt:
  • webgl/2.0.0/conformance2/glsl3/short-circuiting-in-loop-condition-expected.txt:
  • webgl/2.0.0/conformance2/glsl3/tricky-loop-conditions-expected.txt:
  • webgl/2.0.0/conformance2/glsl3/vector-dynamic-indexing-expected.txt:
  • webgl/2.0.0/conformance2/glsl3/vector-dynamic-indexing-nv-driver-bug-expected.txt:
4:59 PM Changeset in webkit [234565] by Ryan Haddad
  • 2 edits in trunk/LayoutTests

Skip imported/w3c/web-platform-tests/FileAPI/reading-data-section/filereader_abort.html.
​https://bugs.webkit.org/show_bug.cgi?id=187269

Unreviewed test gardening.

4:25 PM Changeset in webkit [234564] by commit-queue@webkit.org
  • 62 edits
    2 deletes in trunk/Source/WebCore

Remove WebGPUObject
​https://bugs.webkit.org/show_bug.cgi?id=188114

Patch by Sam Weinig <​sam@webkit.org> on 2018-08-03
Reviewed by Darin Adler.

This removes the unecessary WebGPUObject base class and updates the classes
that were inheriting from it to instead inherit directly from RefCounted. This
shrinks the size of all the classes that were inheriting from it, as they no
longer store a RefPtr<WebGPURenderingContext> member, which is unused. Most
of the classes also shrink because they can be devirtualized.

This also updates most of the create functions and constructors for the classes
that were inheriting from WebGPUObject to no longer require the WebGPURenderingContext
to be passed in.

  • For WebGPUBuffer, WebGPUCommandQueue, WebGPUComputeCommandEncoder, WebGPUComputePipelineState,

WebGPUDepthStencilState, WebGPUDrawable, WebGPUFunction, WebGPULibrary, WebGPURenderCommandEncoder,
WebGPURenderPassDepthAttachmentDescriptor, WebGPURenderPipelineColorAttachmentDescriptor,
WebGPURenderPipelineState, and WebGPUTexture the WebGPURenderingContext and any other additional
parameters needed for initialization were replaced with the underlying type (e.g. WebGPUBuffer
now gets passed a GPUBuffer).

  • WebGPUCommandBuffer is passed a GPUCommandQueue, since it needs to construct the underlying

GPUCommandBuffer itself so it can pass in the completion handler.

  • For WebGPUDepthStencilDescriptor, WebGPURenderPassAttachmentDescriptor,

WebGPURenderPassColorAttachmentDescriptor, WebGPURenderPassDescriptor and WebGPURenderPipelineDescriptor
the create and constructor functions are empty.

  • Sources.txt:
  • WebCore.xcodeproj/project.pbxproj:

Remove WebGPUObject.h/cpp

  • html/canvas/WebGPUBuffer.cpp:

(WebCore::WebGPUBuffer::WebGPUBuffer):

  • html/canvas/WebGPUBuffer.h:
  • html/canvas/WebGPUBuffer.idl:

Switch from inheriting from WebGPUObject to directly inheriting from
RefCounted<WebGPUBuffer>. Since this de-virtualizes it, also add
ImplementationLacksVTable to the IDL file. Also update create
and constructor to take a GPUBuffer directly, rather than the context.

  • html/canvas/WebGPUCommandBuffer.cpp:

(WebCore::WebGPUCommandBuffer::create):
(WebCore::WebGPUCommandBuffer::WebGPUCommandBuffer):
Update create and constructor to no longer require a context, as it would
now be unused, as it no longer inherits from WebGPUObject nor needs to hold
onto the context for any reason.

(WebCore::WebGPUCommandBuffer::createRenderCommandEncoderWithDescriptor):
Update to call the new WebGPURenderCommandEncoder::create, which no longer
requires a context.

(WebCore::WebGPUCommandBuffer::createComputeCommandEncoder):
Update to call the new WebGPUComputeCommandEncoder::create, which no longer
requires a context.

  • html/canvas/WebGPUCommandBuffer.h:
  • html/canvas/WebGPUCommandBuffer.idl:

Switch from inheriting from GPUObject to directly inheriting from
RefCounted<WebGPUCommandBuffer>. Since this de-virtualizes it, also add
ImplementationLacksVTable to the IDL file. Also update create and constructor
to no longer take a context, and just take a GPUCommandQueue. We can't directly
pass a GPUCommandBuffer as WebGPUCommandBuffer needs to pass in a completion handler
on construction of the underlying GPUCommandBuffer.

  • html/canvas/WebGPUCommandQueue.cpp:

(WebCore::WebGPUCommandQueue::WebGPUCommandQueue):
(WebCore::WebGPUCommandQueue::createCommandBuffer):
Update to call the new WebGPUCommandBuffer::create, which no longer
requires a context.

  • html/canvas/WebGPUCommandQueue.h:
  • html/canvas/WebGPUCommandQueue.idl:

Switch from inheriting from GPUObject to directly inheriting from
RefCounted<WebGPUCommandQueue>. Since this de-virtualizes it, also add
ImplementationLacksVTable to the IDL file. Also update create and
constructor to take a GPUCommandQueue directly.

  • html/canvas/WebGPUComputeCommandEncoder.cpp:

(WebCore::WebGPUComputeCommandEncoder::create):
(WebCore::WebGPUComputeCommandEncoder::WebGPUComputeCommandEncoder):

  • html/canvas/WebGPUComputeCommandEncoder.h:
  • html/canvas/WebGPUComputeCommandEncoder.idl:

Switch from inheriting from GPUObject to directly inheriting from
RefCounted<WebGPUComputeCommandEncoder>. Since this de-virtualizes it,
also add ImplementationLacksVTable to the IDL file. Also update create and
constructor to take a GPUComputeCommandEncoder directly

  • html/canvas/WebGPUComputePipelineState.cpp:

(WebCore::WebGPUComputePipelineState::WebGPUComputePipelineState):

  • html/canvas/WebGPUComputePipelineState.h:
  • html/canvas/WebGPUComputePipelineState.idl:

Switch from inheriting from WebGPUObject to directly inheriting from
RefCounted<WebGPUComputePipelineState>. Since this de-virtualizes it, also add
ImplementationLacksVTable to the IDL file. Also update create and
constructor to take a GPUComputePipelineState directly.

  • html/canvas/WebGPUDepthStencilDescriptor.cpp:

Remove unneeded include of include "WebGPURenderingContext.h".

  • html/canvas/WebGPUDepthStencilDescriptor.h:
  • html/canvas/WebGPUDepthStencilDescriptor.idl:

Switch from inheriting from WebGPUObject to directly inheriting from
RefCounted<WebGPUDepthStencilDescriptor>. Since this de-virtualizes it, also add
ImplementationLacksVTable to the IDL file.

  • html/canvas/WebGPUDepthStencilState.cpp:

(WebCore::WebGPUDepthStencilState::WebGPUDepthStencilState):

  • html/canvas/WebGPUDepthStencilState.h:
  • html/canvas/WebGPUDepthStencilState.idl:

Switch from inheriting from WebGPUObject to directly inheriting from
RefCounted<WebGPUDepthStencilState>. Since this de-virtualizes it, also add
ImplementationLacksVTable to the IDL file. Also update create and
constructor to take a GPUDepthStencilState directly.

  • html/canvas/WebGPUDrawable.cpp:

(WebCore::WebGPUDrawable::WebGPUDrawable):

  • html/canvas/WebGPUDrawable.h:
  • html/canvas/WebGPUDrawable.idl:

Switch from inheriting from WebGPUObject to directly inheriting from
RefCounted<WebGPUDrawable>. Since this de-virtualizes it, also add
ImplementationLacksVTable to the IDL file. Also update create and
constructor to take a GPUDrawable directly.

  • html/canvas/WebGPUFunction.cpp:

(WebCore::WebGPUFunction::create):
(WebCore::WebGPUFunction::WebGPUFunction):

  • html/canvas/WebGPUFunction.h:
  • html/canvas/WebGPUFunction.idl:

Switch from inheriting from GPUObject to directly inheriting from
RefCounted<WebGPUFunction>. Since this de-virtualizes it, also add
ImplementationLacksVTable to the IDL file. Also update create and
constructor to take a GPUFunction directly.

  • html/canvas/WebGPULibrary.cpp:

(WebCore::WebGPULibrary::WebGPULibrary):
(WebCore::WebGPULibrary::functionWithName const):
Update for new WebGPUFunction::create, which no longer requires a
context.

  • html/canvas/WebGPULibrary.h:
  • html/canvas/WebGPULibrary.idl:

Switch from inheriting from WebGPUObject to directly inheriting from
RefCounted<WebGPULibrary>. Since this de-virtualizes it, also add
ImplementationLacksVTable to the IDL file. Also update create and
constructor to take a GPULibrary directly, rather than the context.
Unlike the other create functions that take their underlying type, this
one also continues to need the sourceCode String to be passed, as that
is not available from the underlying GPULibrary and must be stored
seperately.

  • html/canvas/WebGPUObject.cpp: Removed.
  • html/canvas/WebGPUObject.h: Removed.

Deleted.

  • html/canvas/WebGPURenderCommandEncoder.cpp:

(WebCore::WebGPURenderCommandEncoder::create):
(WebCore::WebGPURenderCommandEncoder::WebGPURenderCommandEncoder):

  • html/canvas/WebGPURenderCommandEncoder.h:
  • html/canvas/WebGPURenderCommandEncoder.idl:

Switch from inheriting from GPUObject to directly inheriting from
RefCounted<WebGPURenderCommandEncoder>. Since this de-virtualizes it, also add
ImplementationLacksVTable to the IDL file. Also update create and
constructor to take a GPURenderCommandEncoder directly.

  • html/canvas/WebGPURenderPassAttachmentDescriptor.cpp:

(WebCore::WebGPURenderPassAttachmentDescriptor::WebGPURenderPassAttachmentDescriptor):

  • html/canvas/WebGPURenderPassAttachmentDescriptor.h:

Switch from inheriting from GPUObject to directly inheriting from
RefCounted<WebGPURenderPassAttachmentDescriptor>. Also update create and
constructor to no longer take a context, as it is now not used. This remains virtual
as it is inherited from by WebGPURenderPassColorAttachmentDescriptor and WebGPURenderPassDepthAttachmentDescriptor.

  • html/canvas/WebGPURenderPassColorAttachmentDescriptor.cpp:

(WebCore::WebGPURenderPassColorAttachmentDescriptor::create):
(WebCore::WebGPURenderPassColorAttachmentDescriptor::WebGPURenderPassColorAttachmentDescriptor):

  • html/canvas/WebGPURenderPassColorAttachmentDescriptor.h:

Update create and constructor to no longer take a context, as it is now not used. Declare final
as this is not inherited by anything.

  • html/canvas/WebGPURenderPassDepthAttachmentDescriptor.cpp:

(WebCore::WebGPURenderPassDepthAttachmentDescriptor::create):
(WebCore::WebGPURenderPassDepthAttachmentDescriptor::WebGPURenderPassDepthAttachmentDescriptor):

  • html/canvas/WebGPURenderPassDepthAttachmentDescriptor.h:

Update create and constructor to no longer take a context, as it is now not used. Declare final
as this is not inherited by anything.

  • html/canvas/WebGPURenderPassDescriptor.cpp:

(WebCore::WebGPURenderPassDescriptor::depthAttachment):
Update for new WebGPURenderPassDepthAttachmentDescriptor::create function which no longer requires
a context.

Semi-interesting note: This function was previously calling "*context()" to pass a WebGPURenderingContext&
to WebGPURenderPassDepthAttachmentDescriptor::create. However, "context()" was always returning null, as
the constructor of WebGPURenderPassDescriptor was not initializing the WebGPUObject with a WebGPURenderingContext.
Now, one might think that this would cause a null-deref, but, because it was being assigned into a
WebGPURenderingContext&, no dereference happens right away (it essentially just passes a pointer) and since
the function it is being passed to does not use it, it never actually gets dereferenced.

(WebCore::WebGPURenderPassDescriptor::colorAttachments):
Update for new WebGPURenderPassColorAttachmentDescriptor::create function which no longer requires
a context.

  • html/canvas/WebGPURenderPassDescriptor.h:
  • html/canvas/WebGPURenderPassDescriptor.idl:

Switch from inheriting from GPUObject to directly inheriting from
RefCounted<WebGPURenderPassDescriptor>. Since this de-virtualizes it, also add
ImplementationLacksVTable to the IDL file.

  • html/canvas/WebGPURenderPipelineColorAttachmentDescriptor.cpp:

(WebCore::WebGPURenderPipelineColorAttachmentDescriptor::create):
(WebCore::WebGPURenderPipelineColorAttachmentDescriptor::WebGPURenderPipelineColorAttachmentDescriptor):

  • html/canvas/WebGPURenderPipelineColorAttachmentDescriptor.h:
  • html/canvas/WebGPURenderPipelineColorAttachmentDescriptor.idl:

Switch from inheriting from GPUObject to directly inheriting from
RefCounted<WebGPURenderPipelineColorAttachmentDescriptor>. Since this de-virtualizes it, also add
ImplementationLacksVTable to the IDL file. Also update create and constructor to
no longer take a context, as it is now not used.

  • html/canvas/WebGPURenderPipelineDescriptor.cpp:

(WebCore::WebGPURenderPipelineDescriptor::colorAttachments):
Update for new WebGPURenderPipelineColorAttachmentDescriptor::create which no longer requires
a context.

  • html/canvas/WebGPURenderPipelineDescriptor.h:
  • html/canvas/WebGPURenderPipelineDescriptor.idl:

Switch from inheriting from GPUObject to directly inheriting from
RefCounted<WebGPURenderPipelineDescriptor>. Since this de-virtualizes it, also add
ImplementationLacksVTable to the IDL file.

  • html/canvas/WebGPURenderPipelineState.cpp:

(WebCore::WebGPURenderPipelineState::WebGPURenderPipelineState):

  • html/canvas/WebGPURenderPipelineState.h:
  • html/canvas/WebGPURenderPipelineState.idl:

Switch from inheriting from GPUObject to directly inheriting from
RefCounted<WebGPURenderPipelineState>. Since this de-virtualizes it, also add
ImplementationLacksVTable to the IDL file. Also update create and
constructor to take a GPURenderPipelineState directly.

  • html/canvas/WebGPURenderingContext.cpp:

(WebCore::WebGPURenderingContext::createLibrary):
(WebCore::WebGPURenderingContext::createRenderPipelineState):
(WebCore::WebGPURenderingContext::createDepthStencilState):
(WebCore::WebGPURenderingContext::createComputePipelineState):
(WebCore::WebGPURenderingContext::createCommandQueue):
(WebCore::WebGPURenderingContext::nextDrawable):
(WebCore::WebGPURenderingContext::createBuffer):
(WebCore::WebGPURenderingContext::createTexture):
Update for new create functions which require the caller to pass
the underlying type.

  • html/canvas/WebGPUTexture.cpp:

(WebCore::WebGPUTexture::createFromDrawableTexture):
Renamed to create.

(WebCore::WebGPUTexture::create): Deleted
All callers now pass a GPUTexture directly.

  • html/canvas/WebGPUTexture.h:
  • html/canvas/WebGPUTexture.idl:

Switch from inheriting from GPUObject to directly inheriting from
RefCounted<WebGPUTexture>. Since this de-virtualizes it, also add
ImplementationLacksVTable to the IDL file.

  • html/canvas/WebGPUTextureDescriptor.h:
  • html/canvas/WebGPUTextureDescriptor.idl:

Switch from inheriting from GPUObject to directly inheriting from
RefCounted<WebGPUTextureDescriptor>. Since this de-virtualizes it, also add
ImplementationLacksVTable to the IDL file.

4:13 PM Changeset in webkit [234563] by commit-queue@webkit.org
  • 39 edits in trunk/Source

Fix spelling of "overridden"
​https://bugs.webkit.org/show_bug.cgi?id=188315

Patch by Alex Christensen <​achristensen@webkit.org> on 2018-08-03
Reviewed by Darin Adler.

Source/JavaScriptCore:

  • API/JSExport.h:
  • inspector/InjectedScriptSource.js:

Source/ThirdParty/ANGLE:

  • src/libANGLE/renderer/ProgramImpl.h:

Source/ThirdParty/libwebrtc:

  • Source/webrtc/p2p/client/basicportallocator.h:

Source/WebCore:

  • accessibility/ios/WebAccessibilityObjectWrapperIOS.mm:

(-[WebAccessibilityObjectWrapper accessibilityLabel]):
(-[WebAccessibilityObjectWrapper accessibilityValue]):

  • contentextensions/ContentExtensionsBackend.h:
  • editing/EditingStyle.cpp:

(WebCore::EditingStyle::removeStyleFromRulesAndContext):

  • inspector/CommandLineAPIModuleSource.js:
  • platform/ScrollView.h:
  • platform/ScrollableArea.h:
  • platform/mediasession/MediaSessionInterruptionProvider.h:
  • rendering/RenderFragmentedFlow.cpp:

(WebCore::RenderFragmentedFlow::adjustedPositionRelativeToOffsetParent const):

  • rendering/RenderLayer.h:
  • rendering/RenderListBox.h:
  • rendering/svg/RenderSVGResourceClipper.cpp:

(WebCore::RenderSVGResourceClipper::drawContentIntoMaskImage):

  • svg/SVGElement.cpp:

(WebCore::SVGElement::localCoordinateSpaceTransform const):

  • svg/SVGLengthContext.cpp:

(WebCore::SVGLengthContext::SVGLengthContext):
(WebCore::SVGLengthContext::convertValueToUserUnits const):
(WebCore::SVGLengthContext::determineViewport const):

  • svg/SVGLengthContext.h:

Source/WebInspectorUI:

  • UserInterface/Views/ContentBrowserTabContentView.js:

(WI.ContentBrowserTabContentView.prototype.treeElementForRepresentedObject):

  • UserInterface/Views/GeneralTreeElement.js:

(WI.GeneralTreeElement.prototype.ondetach):

  • UserInterface/Views/NavigationBar.js:

(WI.NavigationBar.prototype.get sizesToFit):

  • UserInterface/Views/NavigationSidebarPanel.js:

(WI.NavigationSidebarPanel.prototype.shouldFilterPopulate):

  • UserInterface/Views/TextEditor.js:

(WI.TextEditor.prototype.canBeFormatted):

  • UserInterface/Views/TreeElement.js:

(WI.TreeElement.prototype.onpopulate):

Source/WebKit:

  • Platform/IPC/mac/ConnectionMac.mm:

(IPC::Connection::receiveSourceEventHandler):

  • Shared/WebPreferencesStore.cpp:

(WebKit::WebPreferencesStore::encode const):
(WebKit::WebPreferencesStore::decode):
(WebKit::valueForKey):
(WebKit::setValueForKey):
(WebKit::WebPreferencesStore::setStringValueForKey):
(WebKit::WebPreferencesStore::getStringValueForKey const):
(WebKit::WebPreferencesStore::setBoolValueForKey):
(WebKit::WebPreferencesStore::getBoolValueForKey const):
(WebKit::WebPreferencesStore::setUInt32ValueForKey):
(WebKit::WebPreferencesStore::getUInt32ValueForKey const):
(WebKit::WebPreferencesStore::setDoubleValueForKey):
(WebKit::WebPreferencesStore::getDoubleValueForKey const):
(WebKit::WebPreferencesStore::setOverrideDefaultsStringValueForKey):
(WebKit::WebPreferencesStore::setOverrideDefaultsBoolValueForKey):
(WebKit::WebPreferencesStore::setOverrideDefaultsUInt32ValueForKey):
(WebKit::WebPreferencesStore::setOverrideDefaultsDoubleValueForKey):

  • Shared/WebPreferencesStore.h:
  • UIProcess/ios/WKContentViewInteraction.mm:

(-[WKContentView dragInteraction:previewForLiftingItem:session:]):
(-[WKContentView dragInteraction:previewForCancellingItem:withDefault:]):

  • UIProcess/ios/WKScrollView.mm:

(-[WKScrollView _systemContentInset]):

Source/WebKitLegacy/mac:

  • Misc/WebKitVersionChecks.mm:

(setWebKitLinkTimeVersion):
(WebKitLinkTimeVersion):

  • Plugins/WebBaseNetscapePluginView.mm:

(-[WebBaseNetscapePluginView handleMouseMoved:]):

  • WebCoreSupport/PopupMenuMac.mm:

(PopupMenuMac::populate):

4:10 PM Changeset in webkit [234562] by Kocsen Chung
  • 1 copy in tags/Safari-606.1.36

Tag Safari-606.1.36.

4:01 PM Changeset in webkit [234561] by david_fenton@apple.com
  • 4 edits in trunk

Unreviewed, rolling out r234517.

Caused API test failures on iOS

Reverted changeset:

"[Cocoa] setCookie API fails to set session cookies for
defaultDataStore if processPool created but not used"
​https://bugs.webkit.org/show_bug.cgi?id=188209
​https://trac.webkit.org/changeset/234517

3:50 PM Changeset in webkit [234560] by commit-queue@webkit.org
  • 16 edits
    1 copy in trunk

Add configuration for automatic process pre-warming
​https://bugs.webkit.org/show_bug.cgi?id=187108

Patch by Ben Richards <​benton_richards@apple.com> on 2018-08-03
Reviewed by Ryosuke Niwa.

Source/WebKit:

Added configurations to allow setting the maximum number of processes that should be automatically prewarmed.

  • UIProcess/API/APIProcessPoolConfiguration.cpp:

(API::ProcessPoolConfiguration::copy):

  • UIProcess/API/APIProcessPoolConfiguration.h:
  • UIProcess/API/C/WKContext.cpp:

(WKContextSetMaximumNumberOfPrewarmedProcesses):

  • UIProcess/API/C/WKContextPrivate.h:
  • UIProcess/API/Cocoa/WKProcessPool.mm:

(-[WKProcessPool _setMaximumNumberOfPrewarmedProcesses:]):

  • UIProcess/API/Cocoa/WKProcessPoolPrivate.h:
  • UIProcess/API/Cocoa/_WKProcessPoolConfiguration.h:
  • UIProcess/API/Cocoa/_WKProcessPoolConfiguration.mm:

(-[_WKProcessPoolConfiguration setMaximumPrewarmedProcessCount:]):
(-[_WKProcessPoolConfiguration maximumPrewarmedProcessCount]):

  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::didFinishLoadForFrame): Moved call to notifyProcessPoolToPrewarm from didFirstVisuallyNonEmptyLayoutForFrame to here.
This is to try to ensure that frame loading and prewarming don't happen at the same time as this would be heavy for some devices.
(WebKit::WebPageProxy::didFirstVisuallyNonEmptyLayoutForFrame):

  • UIProcess/WebProcessPool.cpp:

(WebKit::WebProcessPool::setMaximumNumberOfProcesses): Condition changed so that calling setMaximumNumberOfProcesses after warmInitialProcess
doesn't result in a crash.
(WebKit::WebProcessPool::setMaximumNumberOfPrewarmedProcesses):
(WebKit::WebProcessPool::warmInitialProcess):
(WebKit::WebProcessPool::didReachGoodTimeToPrewarm):

  • UIProcess/WebProcessPool.h:

Tools:

Added new test case for setting maximum prewarmed process count and updated test case for process swap on navigation to set the maximum prewarmed process count to be 1 where relevant.

  • TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
  • TestWebKitAPI/Tests/WebKitCocoa/InitialWarmedProcessUsed.mm:

(TEST):

  • TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:
  • TestWebKitAPI/Tests/WebKitCocoa/SetMaximumPrewarmedProcessCount.mm:

(TEST):

2:35 PM Changeset in webkit [234559] by aakash_jain@apple.com
  • 3 edits in trunk/Tools

[ews-build] Add build step to Check Patch Relevance
​https://bugs.webkit.org/show_bug.cgi?id=188295

Reviewed by Lucas Forschler.

  • BuildSlaveSupport/ews-build/steps.py:

(CheckPatchRelevance): Added step to check patch relevance.
(CheckPatchRelevance._patch_is_relevant): Checks if the patch is relevant.
(CheckPatchRelevance._get_patch): Retrieves the patch from buildbot.
(CheckPatchRelevance._addToLog): Add the log message.
(CheckPatchRelevance.start):

  • BuildSlaveSupport/ews-build/factories.py: Added CheckPatchRelevance step appropriately.
2:10 PM Changeset in webkit [234558] by Basuke Suzuki
  • 3 edits in trunk/LayoutTests

[LayoutTests][websocket] Return correct value even when cookie is not set.
​https://bugs.webkit.org/show_bug.cgi?id=188317

Reviewed by Darin Adler.

ws://127.0.0.1:8880/websocket/tests/hybi/echo-cookie_wsh.py

This web socket script throw exception when cookie is not set. That is ended up with
immediate disconnection and causes timeout for failure tests which makes debugging hard.

  • http/tests/websocket/tests/hybi/echo-cookie_wsh.py:

(web_socket_transfer_data):

  • platform/wincairo/TestExpectations:
12:19 PM Changeset in webkit [234557] by Basuke Suzuki
  • 2 edits in trunk/LayoutTests

[Curl] Test gardening
​https://bugs.webkit.org/show_bug.cgi?id=188316

Unreviewed test gardening.

Skip some flaky tests.
Add comments why we skip for some platform dependent reason.

  • platform/wincairo/TestExpectations:
12:07 PM Changeset in webkit [234556] by Kocsen Chung
  • 2 edits in branches/safari-606-branch/Tools

Cherry-pick r234547. rdar://problem/42902608

Unreviewed gardening; r234534 changed the policy for what audio session is set when a
media element is paused. Updating the test to reflect the new policy.

  • TestWebKitAPI/Tests/WebKitLegacy/ios/AudioSessionCategoryIOS.mm: (TestWebKitAPI::TEST):

git-svn-id: ​https://svn.webkit.org/repository/webkit/trunk@234547 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11:36 AM Changeset in webkit [234555] by Ms2ger@igalia.com
  • 2 edits in trunk/Source/WebDriver

Fix invalid optional access in WebDriver Session::createTopLevelBrowsingContext.
​https://bugs.webkit.org/show_bug.cgi?id=188304

Reviewed by Brian Burg.

  • Session.cpp:

(WebDriver::Session::createTopLevelBrowsingContext): remove incorrect value() call.

11:00 AM Changeset in webkit [234554] by Alan Bujtas
  • 3 edits in trunk/Source/WebCore

[LFC][BFC] Special case the document renderer when computing the height and margin
​https://bugs.webkit.org/show_bug.cgi?id=188313

Reviewed by Antti Koivisto.

The document renderer is certainly special and its geometry is likely to be computed as part of "10.6.6 Complicated cases" (spec is unclear) but probably not by
making the it a formatting root. Let's just special case it at contentHeightForFormattingContextRoot(), since having the document renderer as a context root has unwanted side effect.

  • layout/FormattingContextGeometry.cpp:

(WebCore::Layout::contentHeightForFormattingContextRoot):

  • layout/layouttree/LayoutBox.cpp:

(WebCore::Layout::Box::establishesBlockFormattingContext const):

10:57 AM Changeset in webkit [234553] by Alan Bujtas
  • 3 edits in trunk/Source/WebCore

[LFC][BFC] Layout out-of-flow descendants only on the formatting context root.
​https://bugs.webkit.org/show_bug.cgi?id=188312

Reviewed by Antti Koivisto.

Collect out-of-flow descendants on the formatting root container instead of the containing block and run layout on them on the root level. It's more efficient this way.

  • layout/blockformatting/BlockFormattingContext.cpp:

(WebCore::Layout::BlockFormattingContext::layout const):

  • layout/layouttree/LayoutTreeBuilder.cpp:

(WebCore::Layout::TreeBuilder::createSubTree):

10:56 AM Changeset in webkit [234552] by achristensen@apple.com
  • 11 edits
    2 adds in trunk/Source/WebKit

Check with SafeBrowsing during navigation in WKWebView
​https://bugs.webkit.org/show_bug.cgi?id=188133

Reviewed by Chris Dumez.

This turns WebFramePolicyListenerProxy into an object that now listens for the results
of two processes happening in parallel: the API::NavigationClient's decidePolicyForNavigation{Action, Response}
(which it was already waiting for) and, on platforms that support it, the SafariSafeBrowsing framework's check.
The first result is stored as it waits for the second result unless the first result is the API::NavigationClient
saying to cancel or convert the navigation to a download, in which cases we don't care what the safe browsing
framework results are because we won't show the URL in the browser.

Nothing is done with the safe browsing results yet.

  • UIProcess/Cocoa/SafeBrowsingResultCocoa.mm: Added.

(WebKit::SafeBrowsingResult::SafeBrowsingResult):

  • UIProcess/Cocoa/WebPageProxyCocoa.mm:

(WebKit::WebPageProxy::beginSafeBrowsingCheck):

  • UIProcess/SafeBrowsingResult.h: Added.

(WebKit::SafeBrowsingResult::provider const):
(WebKit::SafeBrowsingResult::isPhishing const):
(WebKit::SafeBrowsingResult::isMalware const):
(WebKit::SafeBrowsingResult::isUnwantedSoftware const):
(WebKit::SafeBrowsingResult::isKnownToBeUnsafe const):

  • UIProcess/WebFramePolicyListenerProxy.cpp:

(WebKit::WebFramePolicyListenerProxy::WebFramePolicyListenerProxy):
(WebKit::WebFramePolicyListenerProxy::didReceiveSafeBrowsingResults):
(WebKit::WebFramePolicyListenerProxy::use):
(WebKit::WebFramePolicyListenerProxy::download):
(WebKit::WebFramePolicyListenerProxy::ignore):

  • UIProcess/WebFramePolicyListenerProxy.h:

(WebKit::WebFramePolicyListenerProxy::create):

  • UIProcess/WebFrameProxy.cpp:

(WebKit::WebFrameProxy::setUpPolicyListenerProxy):

  • UIProcess/WebFrameProxy.h:
  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::decidePolicyForNavigationAction):
(WebKit::WebPageProxy::decidePolicyForNewWindowAction):
(WebKit::WebPageProxy::decidePolicyForResponse):

  • UIProcess/WebPageProxy.h:
  • WebKit.xcodeproj/project.pbxproj:
10:51 AM Changeset in webkit [234551] by Basuke Suzuki
  • 2 edits in trunk/LayoutTests

[LayoutTests][PHP] Fix old-style script which uses function name as a keyword.
​https://bugs.webkit.org/show_bug.cgi?id=188289

Reviewed by Alex Christensen.

Function name should be passed as a string literal. In PHP, named function is
not the first class citizen.

  • http/tests/cookies/resources/echo-cookies.php:
10:51 AM Changeset in webkit [234550] by aestes@apple.com
  • 2 edits in trunk/Source/WebCore

ContentFilterUnblockHandler's constructors access moved-from variables
​https://bugs.webkit.org/show_bug.cgi?id=188308
<rdar://problem/42882758>

Reviewed by Alex Christensen.

  • platform/cocoa/ContentFilterUnblockHandlerCocoa.mm:
10:48 AM Changeset in webkit [234549] by Basuke Suzuki
  • 2 edits in trunk/LayoutTests

[LayoutTests] Suppress PHP Error message by removing an execution time limit.
​https://bugs.webkit.org/show_bug.cgi?id=188288

Reviewed by Alex Christensen.

Remove execution timeout.

  • http/tests/multipart/resources/multipart.php:
10:26 AM Changeset in webkit [234548] by Kocsen Chung
  • 13 edits
    3 adds in branches/safari-606-branch

Cherry-pick r234534. rdar://problem/42902608

Control center controls disappear when pausing, locking device.
​https://bugs.webkit.org/show_bug.cgi?id=188282
<rdar://problem/42497809>

Reviewed by Eric Carlson.

Source/WebCore:

Test: platform/mac/media/audio-session-category-video-paused.html

MediaRemote will set any app whose audio session category changes from MediaPlayback to
Ambient as not eligable for being the "now playing" app. Previously, due to the ordering of
events, we never moved from MediaPlayback to Ambient when pausing <video>, even though that
was the intention. Now that that bug is fixed, it exposed this new issue with MediaRemote.

To solve the new issue, make our audio session category policy more explicit: once we are in
MediaPlayback category, we will remain so as long as the media element in queston stays
loaded, has played, and is never interrupted by the system.

Make MediaSessionManagerCocoa a proper subclass of PlatformMediaSessionManager, and make
MediaSessionManageriOS and -Mac subclasses of the -Cocoa class. Add a new
m_hasPlayedSinceLastInterruption ivar to PlatformMediaSession, set when the state() changes
to Playing, and cleared during an interruption. Check this flag when deciding what audio
sessino category to set in MediaSessionManagerCocoa.

  • WebCore.xcodeproj/project.pbxproj:
  • platform/audio/PlatformMediaSession.cpp: (WebCore::PlatformMediaSession::setState):
  • platform/audio/PlatformMediaSession.h: (WebCore::PlatformMediaSession::hasPlayedSinceLastInterruption const): (WebCore::PlatformMediaSession::clearHasPlayedSinceLastInterruption):
  • platform/audio/PlatformMediaSessionManager.cpp: (WebCore::PlatformMediaSessionManager::updateSessionState): Deleted.
  • platform/audio/PlatformMediaSessionManager.h: (WebCore::PlatformMediaSessionManager::audioHardwareListener): (WebCore::PlatformMediaSessionManager::updateSessionState):
  • platform/audio/cocoa/MediaSessionManagerCocoa.cpp: (MediaSessionManagerCocoa::updateSessionState): (MediaSessionManagerCocoa::beginInterruption): (PlatformMediaSessionManager::updateSessionState): Deleted.
  • platform/audio/cocoa/MediaSessionManagerCocoa.h: Added.
  • platform/audio/ios/MediaSessionManagerIOS.h:
  • platform/audio/ios/MediaSessionManagerIOS.mm: (WebCore::MediaSessionManageriOS::MediaSessionManageriOS):
  • platform/audio/mac/MediaSessionManagerMac.h:
  • platform/audio/mac/MediaSessionManagerMac.mm: (WebCore::MediaSessionManagerMac::MediaSessionManagerMac):

LayoutTests:

  • platform/mac/media/audio-session-category-video-paused-expected.txt: Added.
  • platform/mac/media/audio-session-category-video-paused.html: Added.
  • platform/ios/TestExpectations:

git-svn-id: ​https://svn.webkit.org/repository/webkit/trunk@234534 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10:06 AM Changeset in webkit [234547] by jer.noble@apple.com
  • 2 edits in trunk/Tools

Unreviewed gardening; r234534 changed the policy for what audio session is set when a
media element is paused. Updating the test to reflect the new policy.

  • TestWebKitAPI/Tests/WebKitLegacy/ios/AudioSessionCategoryIOS.mm:

(TestWebKitAPI::TEST):

8:38 AM Changeset in webkit [234546] by Alan Bujtas
  • 4 edits in trunk

[LFC][Floating] Now that the document renderer belongs to "complicated cases", adjust viewport stretching.
​https://bugs.webkit.org/show_bug.cgi?id=188306

Reviewed by Antti Koivisto.

Source/WebCore:

And add passing cases for floats.

  • layout/blockformatting/BlockFormattingContextGeometry.cpp:

(WebCore::Layout::initialContainingBlock):
(WebCore::Layout::isStretchedToInitialContainingBlock):
(WebCore::Layout::stretchHeightToInitialContainingBlock):
(WebCore::Layout::stretchWidthToInitialContainingBlock):
(WebCore::Layout::BlockFormattingContext::Geometry::inFlowNonReplacedHeightAndMargin):
(WebCore::Layout::BlockFormattingContext::Geometry::inFlowNonReplacedWidthAndMargin):
(WebCore::Layout::BlockFormattingContext::Geometry::inFlowHeightAndMargin):
(WebCore::Layout::isStretchedToViewport): Deleted.

Tools:

  • LayoutReloaded/misc/LFC-passing-tests.txt:
8:26 AM Changeset in webkit [234545] by Ryan Haddad
  • 11 edits
    2 deletes in trunk/Source/WebKit

Unreviewed, rolling out r234513.

14 API tests fail the assertions added in this change.

Reverted changeset:

"Check with SafeBrowsing during navigation in WKWebView"
​https://bugs.webkit.org/show_bug.cgi?id=188133
​https://trac.webkit.org/changeset/234513

7:21 AM Changeset in webkit [234544] by Alan Bujtas
  • 2 edits in trunk/Source/WebCore

[LFC] Do not check margin box while validating geometry.
​https://bugs.webkit.org/show_bug.cgi?id=188297

Reviewed by Antti Koivisto.

The current RenderBox::marginBox implementation is not spec compliant. Ignore it for now.

  • layout/Verification.cpp:

(WebCore::Layout::outputMismatchingBlockBoxInformationIfNeeded):

3:43 AM Changeset in webkit [234543] by Carlos Garcia Campos
  • 8 edits
    2 adds in trunk/Source/WebKit

[WPE] WebDriver: add support for action commands
​https://bugs.webkit.org/show_bug.cgi?id=188301

Reviewed by Žan Doberšek.

WPE doesn't support action commands because the platform specific code for handling events is not implemented.

  • SourcesWPE.txt: Add new file to compilation.
  • UIProcess/API/glib/WebKitUIClient.cpp: Use the drawing area size as window size in WPE.
  • UIProcess/API/wpe/PageClientImpl.cpp:

(WebKit::PageClientImpl::viewBackend): Return the WPE backend of the view.

  • UIProcess/API/wpe/PageClientImpl.h:
  • UIProcess/Automation/WebAutomationSession.cpp:

(WebKit::WebAutomationSession::performMouseInteraction):
(WebKit::WebAutomationSession::performKeyboardInteractions):
(WebKit::WebAutomationSession::performInteractionSequence):
(WebKit::WebAutomationSession::cancelInteractionSequence):

  • UIProcess/Automation/wpe/WebAutomationSessionWPE.cpp: Added.

(WebKit::modifiersToEventState):
(WebKit::mouseButtonToWPEButton):
(WebKit::stateModifierForWPEButton):
(WebKit::doMouseEvent):
(WebKit::doMotionEvent):
(WebKit::WebAutomationSession::platformSimulateMouseInteraction):
(WebKit::doKeyStrokeEvent):
(WebKit::keyCodeForVirtualKey):
(WebKit::modifiersForKeyCode):
(WebKit::WebAutomationSession::platformSimulateKeyboardInteraction):
(WebKit::WebAutomationSession::platformSimulateKeySequence):

  • UIProcess/WebPageProxy.h:
  • UIProcess/wpe/WebPageProxyWPE.cpp:

(WebKit::WebPageProxy::viewBackend): Return the WPE backend.

3:07 AM Changeset in webkit [234542] by Carlos Garcia Campos
  • 4 edits in trunk/Tools

[WPE] Use the new key mapper API from WPEBackend
​https://bugs.webkit.org/show_bug.cgi?id=188216

Reviewed by Žan Doberšek.

  • WebKitTestRunner/wpe/EventSenderProxyWPE.cpp:

(WTR::EventSenderProxy::keyDown): Use wpe_input_xkb_context_get_entries_for_key_code() to get the hardware key
code.

  • wpe/backends/WindowViewBackend.cpp:

(WPEToolingBackends::WindowViewBackend::WindowViewBackend):
(WPEToolingBackends::WindowViewBackend::handleKeyEvent):
(WPEToolingBackends::WindowViewBackend::modifiers const):

  • wpe/backends/WindowViewBackend.h:
2:38 AM Changeset in webkit [234541] by Carlos Garcia Campos
  • 8 edits in trunk

[WPE] Implement MouseEvent.buttons
​https://bugs.webkit.org/show_bug.cgi?id=187998

Reviewed by Žan Doberšek.

Source/WebKit:

Pass buttons currently pressed to WebMouseEvent.

  • Shared/wpe/WebEventFactory.cpp:

(WebKit::pressedMouseButtons): Helper to get the pressed buttons mask for the WPE modifiers.
(WebKit::WebEventFactory::createWebMouseEvent): Use pressedMouseButtons().

  • UIProcess/API/wpe/PageClientImpl.cpp:

(WebKit::PageClientImpl::doneWithTouchEvent): Update the event modifiers.

Tools:

Pass modifiers to mouse events.

  • WebKitTestRunner/EventSenderProxy.h:
  • WebKitTestRunner/wpe/EventSenderProxyWPE.cpp:

(WTR::senderButtonToWPEButton):
(WTR::modifierForButton):
(WTR::EventSenderProxy::mouseDown):
(WTR::EventSenderProxy::mouseUp):
(WTR::EventSenderProxy::mouseMoveTo):
(WTR::EventSenderProxy::mouseScrollBy):
(WTR::wkEventModifiersToWPE):
(WTR::wpeKeySymForKeyRef):
(WTR::EventSenderProxy::keyDown):
(WTR::EventSenderProxy::prepareAndDispatchTouchEvent):

  • wpe/backends/WindowViewBackend.cpp:

(WPEToolingBackends::WindowViewBackend::handleKeyEvent):
(WPEToolingBackends::WindowViewBackend::modifiers const):

  • wpe/backends/WindowViewBackend.h:
2:16 AM Changeset in webkit [234540] by Carlos Garcia Campos
  • 14 edits in trunk

[WPE] Use WPE key symbols and new API instead of xkbcommon and the key mapper
​https://bugs.webkit.org/show_bug.cgi?id=188093

Reviewed by Žan Doberšek.

Source/WebCore:

Add helper functions to get key code, hardware key code, key identifier and windows key code from a WPE key symbol.

  • platform/PlatformKeyboardEvent.h:
  • platform/wpe/PlatformKeyboardEventWPE.cpp:

(WebCore::PlatformKeyboardEvent::keyValueForWPEKeyCode):
(WebCore::PlatformKeyboardEvent::keyCodeForHardwareKeyCode):
(WebCore::PlatformKeyboardEvent::keyIdentifierForWPEKeyCode):
(WebCore::PlatformKeyboardEvent::windowsKeyCodeForWPEKeyCode):
(WebCore::PlatformKeyboardEvent::singleCharacterString):

Source/WebKit:

  • Shared/WebEvent.h: Add WebKeyboardEvent constructor for WPE.
  • Shared/WebKeyboardEvent.cpp:

(WebKit::WebKeyboardEvent::WebKeyboardEvent): WebKeyboardEvent constructor for WPE.

  • Shared/wpe/WebEventFactory.cpp:

(WebKit::isWPEKeyCodeFromKeyPad): Helper tpo check if a key symbols is from key pad.
(WebKit::WebEventFactory::createWebKeyboardEvent): Use PlatformKeyboardEvent API to provide key code, hardware
key code, key identifier and windows key code to WebKeyboardEvent.

  • UIProcess/API/wpe/WPEView.cpp:

(WKWPE::m_backend): Use WPE key symbols.

Tools:

  • MiniBrowser/wpe/main.cpp: Use WPE key symbols.
  • WebKitTestRunner/wpe/EventSenderProxyWPE.cpp:

(WTR::wpeKeySymForKeyRef): Use WPE key symbols and wpe_unicode_to_key_code().
(WTR::EventSenderProxy::keyDown): Do not set unicode char to the event since it was removed.

  • flatpak/org.webkit.WPE.yaml: Bump WPEBackend and WPEBackend-fdo versions.
  • wpe/backends/WindowViewBackend.cpp:

(WPEToolingBackends::WindowViewBackend::handleKeyEvent): Do not set unicode char to the event since it was removed.

  • wpe/jhbuild.modules: Bump WPEBackend and WPEBackend-fdo versions.
1:16 AM Changeset in webkit [234539] by rniwa@webkit.org
  • 10 edits
    1 add in trunk

Release assert when throwing exceptions in custom element reactions
​https://bugs.webkit.org/show_bug.cgi?id=187805
<rdar://problem/42432714>

Reviewed by Saam Barati.

LayoutTests/imported/w3c:

Generated the expected result.

  • web-platform-tests/custom-elements/reactions/with-exceptions-expected.txt: Added.

Source/WebCore:

The release assertion was hit because we were not catching & re-throwing the exception thrown by DOM API
before trying to execute custom elements reactions in ~CustomElementReactionStack as specified here:
​https://html.spec.whatwg.org/multipage/custom-elements.html#cereactions
Fixed the bug by capturing the exception and re-throwing the exception as specified.

Tests: imported/w3c/web-platform-tests/custom-elements/reactions/with-exceptions.html

  • bindings/js/JSMainThreadExecState.h:

(WebCore::JSMainThreadNullState::JSMainThreadNullState): Use the previous JS state.

  • bindings/scripts/CodeGeneratorJS.pm:

(GeneratePut): Pass in the exec state to CustomElementReactionStack.
(GeneratePutByIndex): Ditto.
(GenerateDefineOwnProperty): Ditto.
(GenerateDeletePropertyCommon): Ditto.
(GenerateAttributeSetterBodyDefinition): Ditto.
(GenerateOperationBodyDefinition): Ditto.

  • bindings/scripts/test/JS/JSTestCEReactions.cpp:

(WebCore::setJSTestCEReactionsAttributeWithCEReactionsSetter):
(WebCore::setJSTestCEReactionsReflectAttributeWithCEReactionsSetter):
(WebCore::jsTestCEReactionsPrototypeFunctionMethodWithCEReactionsBody):

  • bindings/scripts/test/JS/JSTestCEReactionsStringifier.cpp:

(WebCore::setJSTestCEReactionsStringifierValueSetter):

  • dom/CustomElementReactionQueue.cpp:

(WebCore::CustomElementReactionQueue::ElementQueue::processQueue): Added. If there is a script running
in the stack (i.e. ExecState is not null), catch any exception before executing custom element reactions,
then re-throw the exception afterwards. ExecState is null when DOM API is invoked via Objective-C bindings
or when custom element reactions are executed in the backup queue (e.g. for editing operations).
(WebCore::CustomElementReactionStack::processQueue):
(WebCore::CustomElementReactionQueue::processBackupQueue):

  • dom/CustomElementReactionQueue.h:

(WebCore::CustomElementReactionStack::CustomElementReactionStack):
(WebCore::CustomElementReactionStack::~CustomElementReactionStack):

LayoutTests:

Unskipped the previously crashing test.

Aug 2, 2018:

10:24 PM Changeset in webkit [234538] by Basuke Suzuki
  • 2 edits
    7 adds in trunk/LayoutTests

[Curl] Test gardening
​https://bugs.webkit.org/show_bug.cgi?id=188293

Unreviewed test gardening.

Unskip some http tests.

  • platform/wincairo/TestExpectations:
    • contentextensions
    • events
    • favicon.ico
    • fullscreen
    • globalhistory
    • gzip-content-encoding
    • history
    • local
    • messaging
    • mime
    • multipart
    • performance
    • permissionclient
    • referrer-policy
    • resourceLoadStatistics
  • platform/wincairo/http/tests/local/file-url-sent-as-referer-expected.txt: Added.
  • platform/wincairo/http/tests/multipart/invalid-image-data-expected.txt: Added.
  • platform/wincairo/http/tests/multipart/invalid-image-data-standalone-expected.txt: Added.
  • platform/wincairo/http/tests/uri/css-href-expected.txt: Added.
9:32 PM Changeset in webkit [234537] by Alan Bujtas
  • 6 edits in trunk/Source/WebCore

[LFC][BFC] Apply the "10.6.6 Complicated cases" when computing height and margin for the document renderer
​https://bugs.webkit.org/show_bug.cgi?id=188296

Reviewed by Simon Fraser.

The spec is not clear about what to do with the document renderer. It behaves as if it was a formatting context root when
it comes to computing the content height. Let's apply "10.6.6 Complicated cases" for now.

  • layout/FormattingContext.h:
  • layout/FormattingContextGeometry.cpp:

(WebCore::Layout::FormattingContext::Geometry::complicatedCases):
(WebCore::Layout::FormattingContext::Geometry::floatingHeightAndMargin):
(WebCore::Layout::FormattingContext::Geometry::floatingNonReplacedHeightAndMargin): Deleted.

  • layout/blockformatting/BlockFormattingContextGeometry.cpp:

(WebCore::Layout::BlockFormattingContext::Geometry::inFlowNonReplacedHeightAndMargin):
(WebCore::Layout::BlockFormattingContext::Geometry::inFlowHeightAndMargin):

  • layout/layouttree/LayoutBox.cpp:

(WebCore::Layout::Box::establishesBlockFormattingContext const):

  • layout/layouttree/LayoutBox.h:
7:32 PM Changeset in webkit [234536] by Alan Bujtas
  • 4 edits in trunk/Source/WebCore

[LFC][Floating] Add FloatingState::bottom() to enable content height computation for formatting roots.
​https://bugs.webkit.org/show_bug.cgi?id=188294

Reviewed by Simon Fraser.

"10.6.7 'Auto' heights for block formatting context roots
...
In addition, if the element has any floating descendants whose bottom margin edge is below the element's bottom content edge,
then the height is increased to include those edges. Only floats that participate in this block formatting context are taken into
account, e.g., floats inside absolutely positioned descendants or other floats are not..."

  • layout/FloatingState.cpp:

(WebCore::Layout::FloatingState::bottom const):

  • layout/FloatingState.h:
  • layout/FormattingContextGeometry.cpp:

(WebCore::Layout::contentHeightForFormattingContextRoot):

7:10 PM Changeset in webkit [234535] by bshafiei@apple.com
  • 9 edits in branches/safari-606-branch/Source

Cherry-pick r234528. rdar://problem/42883788

Source/JavaScriptCore:
Reading instructionPointer from PlatformRegisters may fail when using pointer profiling
​https://bugs.webkit.org/show_bug.cgi?id=188271
<rdar://problem/42850884>

Reviewed by Michael Saboff.

This patch defends against the instructionPointer containing garbage bits.
See radar for details.

  • runtime/MachineContext.h: (JSC::MachineContext::instructionPointer):
  • runtime/SamplingProfiler.cpp: (JSC::SamplingProfiler::takeSample):
  • runtime/VMTraps.cpp: (JSC::SignalContext::SignalContext): (JSC::SignalContext::tryCreate):
  • tools/CodeProfiling.cpp: (JSC::profilingTimer):
  • tools/SigillCrashAnalyzer.cpp: (JSC::SignalContext::SignalContext): (JSC::SignalContext::tryCreate): (JSC::SignalContext::dump): (JSC::installCrashHandler):
  • wasm/WasmFaultSignalHandler.cpp: (JSC::Wasm::trapHandler):

Source/WTF:
Reading instructionPointer from PlatformRegisters may fail when using pointer tagging
​https://bugs.webkit.org/show_bug.cgi?id=188271
<rdar://problem/42850884>

Reviewed by Michael Saboff.

  • wtf/PtrTag.h: (WTF::isTaggedWith): (WTF::usesPointerTagging):

git-svn-id: ​https://svn.webkit.org/repository/webkit/trunk@234528 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6:54 PM Changeset in webkit [234534] by jer.noble@apple.com
  • 13 edits
    3 adds in trunk

Control center controls disappear when pausing, locking device.
​https://bugs.webkit.org/show_bug.cgi?id=188282
<rdar://problem/42497809>

Reviewed by Eric Carlson.

Source/WebCore:

Test: platform/mac/media/audio-session-category-video-paused.html

MediaRemote will set any app whose audio session category changes from MediaPlayback to
Ambient as not eligable for being the "now playing" app. Previously, due to the ordering of
events, we never moved from MediaPlayback to Ambient when pausing <video>, even though that
was the intention. Now that that bug is fixed, it exposed this new issue with MediaRemote.

To solve the new issue, make our audio session category policy more explicit: once we are in
MediaPlayback category, we will remain so as long as the media element in queston stays
loaded, has played, and is never interrupted by the system.

Make MediaSessionManagerCocoa a proper subclass of PlatformMediaSessionManager, and make
MediaSessionManageriOS and -Mac subclasses of the -Cocoa class. Add a new
m_hasPlayedSinceLastInterruption ivar to PlatformMediaSession, set when the state() changes
to Playing, and cleared during an interruption. Check this flag when deciding what audio
sessino category to set in MediaSessionManagerCocoa.

  • WebCore.xcodeproj/project.pbxproj:
  • platform/audio/PlatformMediaSession.cpp:

(WebCore::PlatformMediaSession::setState):

  • platform/audio/PlatformMediaSession.h:

(WebCore::PlatformMediaSession::hasPlayedSinceLastInterruption const):
(WebCore::PlatformMediaSession::clearHasPlayedSinceLastInterruption):

  • platform/audio/PlatformMediaSessionManager.cpp:

(WebCore::PlatformMediaSessionManager::updateSessionState): Deleted.

  • platform/audio/PlatformMediaSessionManager.h:

(WebCore::PlatformMediaSessionManager::audioHardwareListener):
(WebCore::PlatformMediaSessionManager::updateSessionState):

  • platform/audio/cocoa/MediaSessionManagerCocoa.cpp:

(MediaSessionManagerCocoa::updateSessionState):
(MediaSessionManagerCocoa::beginInterruption):
(PlatformMediaSessionManager::updateSessionState): Deleted.

  • platform/audio/cocoa/MediaSessionManagerCocoa.h: Added.
  • platform/audio/ios/MediaSessionManagerIOS.h:
  • platform/audio/ios/MediaSessionManagerIOS.mm:

(WebCore::MediaSessionManageriOS::MediaSessionManageriOS):

  • platform/audio/mac/MediaSessionManagerMac.h:
  • platform/audio/mac/MediaSessionManagerMac.mm:

(WebCore::MediaSessionManagerMac::MediaSessionManagerMac):

LayoutTests:

  • platform/mac/media/audio-session-category-video-paused-expected.txt: Added.
  • platform/mac/media/audio-session-category-video-paused.html: Added.
  • platform/ios/TestExpectations:
6:16 PM Changeset in webkit [234533] by bshafiei@apple.com
  • 7 edits in branches/safari-606-branch/Source/WebCore

Cherry-pick r234512. rdar://problem/42882298

Text selection color is hard to see in dark mode web views.
​https://bugs.webkit.org/show_bug.cgi?id=188260
rdar://problem/42721294

Reviewed by Simon Fraser.

Stop using blendWithWhite() to transform the AppKit selection color in dark mode.
Using an alpha of 80% gives good contrast, and still works good for selections over images.

  • platform/graphics/Color.cpp: (WebCore::Color::blendWithWhite const): Mark new colors as semantic if the original is. (WebCore::Color::colorWithAlpha const): Ditto.
  • rendering/RenderElement.cpp: (WebCore::RenderElement::selectionBackgroundColor const): Use transformSelectionBackgroundColor.
  • rendering/RenderTheme.cpp: (WebCore::RenderTheme::activeSelectionBackgroundColor const): Use transformSelectionBackgroundColor. (WebCore::RenderTheme::inactiveSelectionBackgroundColor const): Ditto. (WebCore::RenderTheme::transformSelectionBackgroundColor const): Added. Just blend with white.
  • rendering/RenderTheme.h:
  • rendering/RenderThemeMac.h:
  • rendering/RenderThemeMac.mm: (WebCore::RenderThemeMac::transformSelectionBackgroundColor const): Added. Use an alpha with the color in dark mode, otherwise fallback to RenderTheme. (WebCore::RenderThemeMac::systemColor const): Use activeListBoxSelectionBackgroundColor() and activeSelectionBackgroundColor() instead of caching the colors again. Update hardcoded color.

git-svn-id: ​https://svn.webkit.org/repository/webkit/trunk@234512 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6:16 PM Changeset in webkit [234532] by bshafiei@apple.com
  • 2 edits in branches/safari-606-branch/Source/WebCore

Cherry-pick r234469. rdar://problem/42882308

Always use MediaPlayback audio category when playing to AppleTV
​https://bugs.webkit.org/show_bug.cgi?id=188230
<rdar://problem/42497809>

Reviewed by Jer Noble.

  • platform/audio/cocoa/MediaSessionManagerCocoa.cpp: (PlatformMediaSessionManager::updateSessionState): Check session.isPlayingToWirelessPlaybackTarget().

git-svn-id: ​https://svn.webkit.org/repository/webkit/trunk@234469 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6:16 PM Changeset in webkit [234531] by bshafiei@apple.com
  • 4 edits in branches/safari-606-branch/Source/WebCore

Cherry-pick r234463. rdar://problem/42882308

[iOS] Remove the delay before setting audio session category added in r233535
​https://bugs.webkit.org/show_bug.cgi?id=188225

Reviewed by Jer Noble.

  • platform/audio/PlatformMediaSessionManager.cpp: (WebCore::PlatformMediaSessionManager::beginInterruption): scheduleUpdateSessionState -> updateSessionState. (WebCore::PlatformMediaSessionManager::addSession): Ditto. (WebCore::PlatformMediaSessionManager::removeSession): Ditto. (WebCore::PlatformMediaSessionManager::sessionStateChanged): Ditto. (WebCore::PlatformMediaSessionManager::sessionCanProduceAudioChanged): Ditto. (WebCore::PlatformMediaSessionManager::updateSessionState): Ditto. (WebCore::PlatformMediaSessionManager::audioOutputDeviceChanged): Ditto. (WebCore::PlatformMediaSessionManager::scheduleUpdateSessionState): Deleted.
  • platform/audio/PlatformMediaSessionManager.h:
  • platform/audio/cocoa/MediaSessionManagerCocoa.cpp: (PlatformMediaSessionManager::updateSessionState): (PlatformMediaSessionManager::scheduleUpdateSessionState): Deleted.
  • rendering/RenderLayer.cpp: (WebCore::RenderLayer::updateClipRects):

git-svn-id: ​https://svn.webkit.org/repository/webkit/trunk@234463 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6:11 PM Changeset in webkit [234530] by timothy_horton@apple.com
  • 4 edits in trunk/Source/WebKit

PDFPlugin: Context menus in RTL are left-aligned
​https://bugs.webkit.org/show_bug.cgi?id=188292
<rdar://problem/32293787>

Reviewed by Simon Fraser.

  • WebProcess/Plugins/PDF/PDFLayerControllerSPI.h:

Add some SPI.

  • WebProcess/Plugins/PDF/PDFPlugin.mm:

(WebKit::PDFPlugin::handleContextMenuEvent):
Translate UI layout direction back into the platform enum, and pass it to PDFKit.

  • WebProcess/WebPage/WebPage.h:

(WebKit::WebPage::userInterfaceLayoutDirection const):
Add a getter for UI layout direction.

5:24 PM Changeset in webkit [234529] by Chris Dumez
  • 2 edits in trunk/Source/WebKit

Regression(r234486): assertion hit in ~CallbackAggregator()
​https://bugs.webkit.org/show_bug.cgi?id=188283
<rdar://problem/42851342>

Reviewed by Alex Christensen.

[NSHTTPCookieStorage _saveCookies:] SPI may call its completion block on the background queue
so we need to make sure we dispatch back to the main thread before calling our completion
handler.

  • NetworkProcess/cocoa/NetworkProcessCocoa.mm:

(WebKit::saveCookies):
(WebKit::NetworkProcess::platformSyncAllCookies):

5:14 PM Changeset in webkit [234528] by sbarati@apple.com
  • 9 edits in trunk/Source

Source/JavaScriptCore:
Reading instructionPointer from PlatformRegisters may fail when using pointer profiling
​https://bugs.webkit.org/show_bug.cgi?id=188271
<rdar://problem/42850884>

Reviewed by Michael Saboff.

This patch defends against the instructionPointer containing garbage bits.
See radar for details.

  • runtime/MachineContext.h:

(JSC::MachineContext::instructionPointer):

  • runtime/SamplingProfiler.cpp:

(JSC::SamplingProfiler::takeSample):

  • runtime/VMTraps.cpp:

(JSC::SignalContext::SignalContext):
(JSC::SignalContext::tryCreate):

  • tools/CodeProfiling.cpp:

(JSC::profilingTimer):

  • tools/SigillCrashAnalyzer.cpp:

(JSC::SignalContext::SignalContext):
(JSC::SignalContext::tryCreate):
(JSC::SignalContext::dump):
(JSC::installCrashHandler):

  • wasm/WasmFaultSignalHandler.cpp:

(JSC::Wasm::trapHandler):

Source/WTF:
Reading instructionPointer from PlatformRegisters may fail when using pointer tagging
​https://bugs.webkit.org/show_bug.cgi?id=188271
<rdar://problem/42850884>

Reviewed by Michael Saboff.

  • wtf/PtrTag.h:

(WTF::isTaggedWith):
(WTF::usesPointerTagging):

5:00 PM Changeset in webkit [234527] by Ryan Haddad
  • 2 edits in branches/safari-606-branch/Tools

Cherry-pick r234521. rdar://problem/42387347

Skip API test DataInteractionTests.CustomActionSheetPopover on current SDKs.
​https://bugs.webkit.org/show_bug.cgi?id=175204

Unreviewed test gardening.

  • TestWebKitAPI/Tests/ios/DataInteractionTests.mm:

git-svn-id: ​http://svn.webkit.org/repository/webkit/trunk@234521 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4:54 PM Changeset in webkit [234526] by n_wang@apple.com
  • 11 edits
    2 adds in trunk

AX: [iOS] add support to return the attributed string under the element
​https://bugs.webkit.org/show_bug.cgi?id=188276
<rdar://problem/42872357>

Reviewed by Chris Fleizach.

Source/WebCore:

Provided a way on iOS to return the attributed string under the element for
better performance.

Test: accessibility/ios-simulator/attributed-string-for-element.html

  • accessibility/ios/WebAccessibilityObjectWrapperIOS.mm:

(-[WebAccessibilityObjectWrapper _stringFromStartMarker:toEndMarker:attributed:]):
(-[WebAccessibilityObjectWrapper _stringForRange:attributed:]):
(-[WebAccessibilityObjectWrapper attributedStringForElement]):

Tools:

  • DumpRenderTree/AccessibilityUIElement.cpp:

(attributedStringForElementCallback):
(AccessibilityUIElement::getJSClass):

  • DumpRenderTree/AccessibilityUIElement.h:
  • DumpRenderTree/ios/AccessibilityUIElementIOS.mm:

(AccessibilityUIElement::attributedStringForElement):

  • WebKitTestRunner/InjectedBundle/AccessibilityUIElement.cpp:

(WTR::AccessibilityUIElement::attributedStringForElement):

  • WebKitTestRunner/InjectedBundle/AccessibilityUIElement.h:
  • WebKitTestRunner/InjectedBundle/Bindings/AccessibilityUIElement.idl:
  • WebKitTestRunner/InjectedBundle/ios/AccessibilityUIElementIOS.mm:

(WTR::AccessibilityUIElement::attributedStringForElement):

LayoutTests:

  • accessibility/ios-simulator/attributed-string-for-element-expected.txt: Added.
  • accessibility/ios-simulator/attributed-string-for-element.html: Added.
4:48 PM Changeset in webkit [234525] by Kocsen Chung
  • 7 edits in branches/safari-606-branch/Source

Versioning.

4:46 PM Changeset in webkit [234524] by Alan Bujtas
  • 2 edits in trunk/Source/WebCore

[LFC][Floating] Remove redundant LayoutContext member variable.
​https://bugs.webkit.org/show_bug.cgi?id=188286

Reviewed by Simon Fraser.

  • layout/FloatingContext.cpp:

(WebCore::Layout::begin):
(WebCore::Layout::end):
(WebCore::Layout::FloatingContext::floatingPosition const):
(WebCore::Layout::FloatingPair::FloatingPair):
(WebCore::Layout::Iterator::Iterator):

4:41 PM Changeset in webkit [234523] by Kocsen Chung
  • 1 copy in tags/Safari-606.1.35

Tag Safari-606.1.35.

4:35 PM Changeset in webkit [234522] by timothy_horton@apple.com
  • 1 edit
    2 deletes in trunk/LayoutTests

REGRESSION (r234488): Layout test fast/events/ios/keyboard-scrolling-repeat.html is timing out
<rdar://problem/42878480>

  • fast/events/ios/keyboard-scrolling-repeat-expected.txt: Removed.
  • fast/events/ios/keyboard-scrolling-repeat.html: Removed.

Remove a test that can't work anymore.
If we need to re-write it it will probably need to be an API test.

4:09 PM Changeset in webkit [234521] by Ryan Haddad
  • 2 edits in trunk/Tools

Skip API test DataInteractionTests.CustomActionSheetPopover on current SDKs.
​https://bugs.webkit.org/show_bug.cgi?id=175204

Unreviewed test gardening.

  • TestWebKitAPI/Tests/ios/DataInteractionTests.mm:
3:37 PM Changeset in webkit [234520] by rniwa@webkit.org
  • 3 edits in trunk/Source/WebCore

Move queue processing logic from CustomElementReactionStack to CustomElementReactionQueue
​https://bugs.webkit.org/show_bug.cgi?id=188277

Reviewed by Wenson Hsieh.

Moved ensureCurrentQueue, processBackupQueue, ElementQueue, ensureBackupQueue, and backupElementQueue
from CustomElementReactionStack to CustomElementReactionQueue to simplify CustomElementReactionStack.

No new tests since there is no behavior change.

  • dom/CustomElementReactionQueue.cpp:

(WebCore::CustomElementReactionQueue::enqueueElementUpgrade):
(WebCore::CustomElementReactionQueue::enqueueConnectedCallbackIfNeeded):
(WebCore::CustomElementReactionQueue::enqueueDisconnectedCallbackIfNeeded):
(WebCore::CustomElementReactionQueue::enqueueAdoptedCallbackIfNeeded):
(WebCore::CustomElementReactionQueue::enqueueAttributeChangedCallbackIfNeeded):
(WebCore::CustomElementReactionQueue::ElementQueue::add): Moved from CustomElementReactionStack.
(WebCore::CustomElementReactionQueue::ElementQueue::invokeAll): Ditto.
(WebCore::CustomElementReactionQueue::ensureCurrentQueue): Ditto.
(WebCore::CustomElementReactionQueue::ensureBackupQueue): Ditto.
(WebCore::CustomElementReactionQueue::processBackupQueue): Ditto.
(WebCore::CustomElementReactionQueue::backupElementQueue): Ditto.

  • dom/CustomElementReactionQueue.h:

(WebCore::CustomElementReactionStack::hasCurrentProcessingStack): Deleted. It was never called.

3:31 PM Changeset in webkit [234519] by achristensen@apple.com
  • 2 edits in trunk/Source/WebKit

Fix some builds after r234516
​​https://bugs.webkit.org/show_bug.cgi?id=188250

  • Platform/IPC/ArgumentCoders.h:

This assertion was comparing integers of different signs.
Since we don't allow exception throwing in WebKit, valueless_by_exception won't be true,
so this assertion will always be true. I'll just remove it.

3:20 PM Changeset in webkit [234518] by sihui_liu@apple.com
  • 3 edits in trunk/Source/WebKit

httpCookieStore.getAllCookies() does not always call completion block
​https://bugs.webkit.org/show_bug.cgi?id=188242

Reviewed by Chris Dumez.

Take background assertion for network process to keep it alive during cookieStore API calls.

  • UIProcess/GenericCallback.h:
  • UIProcess/WebCookieManagerProxy.cpp:

(WebKit::WebCookieManagerProxy::getHostnamesWithCookies):
(WebKit::WebCookieManagerProxy::deleteCookie):
(WebKit::WebCookieManagerProxy::deleteAllCookiesModifiedSince):
(WebKit::WebCookieManagerProxy::setCookie):
(WebKit::WebCookieManagerProxy::setCookies):
(WebKit::WebCookieManagerProxy::getAllCookies):
(WebKit::WebCookieManagerProxy::getCookies):
(WebKit::WebCookieManagerProxy::setHTTPCookieAcceptPolicy):
(WebKit::WebCookieManagerProxy::getHTTPCookieAcceptPolicy):

3:18 PM Changeset in webkit [234517] by sihui_liu@apple.com
  • 4 edits in trunk

[Cocoa] setCookie API fails to set session cookies for defaultDataStore if processPool created but not used
​https://bugs.webkit.org/show_bug.cgi?id=188209

Reviewed by Geoffrey Garen.

Source/WebKit:

If defaultDataStore is already created when launching network process, it means user
probably set cookies for default session via API, so we should update m_websiteDataStore to
make sure pendingCookies is passed to network process.

  • UIProcess/WebProcessPool.cpp:

(WebKit::WebProcessPool::ensureNetworkProcess):

Tools:

  • TestWebKitAPI/Tests/WebKitCocoa/WKHTTPCookieStore.mm:

(TEST):

2:48 PM Changeset in webkit [234516] by achristensen@apple.com
  • 8 edits in trunk/Source/WebKit

Use WTF::Variant for WebPreferencesStore::Value
​https://bugs.webkit.org/show_bug.cgi?id=188250

Reviewed by Sam Weinig.

It was conceptually a variant. This just uses an actual WTF::Variant.
Since it's encoded and decoded, I wrote variadic template encoders and decoders
like we have for other types (Expected, std::tuple, KeyValuePair, etc.)

  • Platform/IPC/ArgumentCoders.h:

(IPC::VariantCoder::encode):
(IPC::VariantCoder::decode):
Introduce ArgumentCoder<Variant>, which encodes which type the Variant currently
holds followed by the encoding of the current type.

  • Platform/IPC/Decoder.cpp:

(IPC::Decoder::Decoder):

  • Platform/IPC/Decoder.h:
  • Platform/IPC/mac/ConnectionMac.mm:

(IPC::createMessageDecoder):
Use move semantics to prevent an unnecessary copy of the Vector<Attachment>

  • Scripts/PreferencesTemplates/WebPreferencesStoreDefaultsMap.cpp.erb:
  • Shared/WebPreferencesStore.cpp:

(WebKit::WebPreferencesStore::decode):
Use modern std::optional-based decoding supported by the new Variant decoding.
(WebKit::valueForKey):
Use WTF::get and WTF::holds_alternative instead of custom type checks.
(WebKit::WebPreferencesStore::Value::encode const): Deleted.
(WebKit::WebPreferencesStore::Value::decode): Deleted.
(WebKit::as<String>): Deleted.
(WebKit::as<bool>): Deleted.
(WebKit::as<uint32_t>): Deleted.
(WebKit::as<double>): Deleted.

  • Shared/WebPreferencesStore.h:

(WebKit::WebPreferencesStore::Value::Value): Deleted.
(WebKit::WebPreferencesStore::Value::operator=): Deleted.
(WebKit::WebPreferencesStore::Value::~Value): Deleted.
(WebKit::WebPreferencesStore::Value::type const): Deleted.
(WebKit::WebPreferencesStore::Value::asString const): Deleted.
(WebKit::WebPreferencesStore::Value::asBool const): Deleted.
(WebKit::WebPreferencesStore::Value::asUInt32 const): Deleted.
(WebKit::WebPreferencesStore::Value::asDouble const): Deleted.
(WebKit::WebPreferencesStore::Value::destroy): Deleted.
(): Deleted.
Use WTF::Variant instead of a custom type/union pair.

2:26 PM Changeset in webkit [234515] by Megan Gardner
  • 9 edits in trunk/LayoutTests

Reformat editable selection tests to remove potential flakeyness due to autoscroll
​https://bugs.webkit.org/show_bug.cgi?id=188263
<rdar://problem/40565375>

Reviewed by Wenson Hsieh

Change the format of these editable selection tests to put the editable content at the top of the page so
that when selecting we do not run into the page trying to autoscroll while selecting, which makes the tests more
flakey, and more difficult to get accurate and consistent answers.

  • fast/events/touch/ios/double-tap-on-editable-content-for-selection-then-drag-down-to-change-selected-text.html:
  • fast/events/touch/ios/double-tap-on-editable-content-for-selection-then-drag-left-to-change-selected-text.html:
  • fast/events/touch/ios/double-tap-on-editable-content-for-selection-then-drag-right-to-change-selected-text.html:
  • fast/events/touch/ios/double-tap-on-editable-content-for-selection-then-drag-up-to-change-selected-text.html:
  • fast/events/touch/ios/long-press-on-editable-content-then-drag-down-to-change-selected-text.html:
  • fast/events/touch/ios/long-press-on-editable-content-then-drag-left-to-change-selected-text.html:
  • fast/events/touch/ios/long-press-on-editable-content-then-drag-right-to-change-selected-text.html:
  • fast/events/touch/ios/long-press-on-editable-content-then-drag-up-to-change-selected-text.html:
1:24 PM Changeset in webkit [234514] by achristensen@apple.com
  • 2 edits in trunk/Source/WTF

Add some Variant types to Forward.h
​https://bugs.webkit.org/show_bug.cgi?id=188250

  • wtf/Forward.h:
1:13 PM Changeset in webkit [234513] by achristensen@apple.com
  • 11 edits
    2 adds in trunk/Source/WebKit

Check with SafeBrowsing during navigation in WKWebView
​https://bugs.webkit.org/show_bug.cgi?id=188133

Reviewed by Chris Dumez.

This turns WebFramePolicyListenerProxy into an object that now listens for the results
of two processes happening in parallel: the API::NavigationClient's decidePolicyForNavigation{Action, Response}
(which it was already waiting for) and, on platforms that support it, the SafariSafeBrowsing framework's check.
The first result is stored as it waits for the second result unless the first result is the API::NavigationClient
saying to cancel or convert the navigation to a download, in which cases we don't care what the safe browsing
framework results are because we won't show the URL in the browser.

Nothing is done with the safe browsing results yet.

  • UIProcess/Cocoa/SafeBrowsingResultCocoa.mm: Added.

(WebKit::SafeBrowsingResult::SafeBrowsingResult):

  • UIProcess/Cocoa/WebPageProxyCocoa.mm:

(WebKit::WebPageProxy::beginSafeBrowsingCheck):

  • UIProcess/SafeBrowsingResult.h: Added.

(WebKit::SafeBrowsingResult::provider const):
(WebKit::SafeBrowsingResult::isPhishing const):
(WebKit::SafeBrowsingResult::isMalware const):
(WebKit::SafeBrowsingResult::isUnwantedSoftware const):
(WebKit::SafeBrowsingResult::isKnownToBeUnsafe const):

  • UIProcess/WebFramePolicyListenerProxy.cpp:

(WebKit::WebFramePolicyListenerProxy::WebFramePolicyListenerProxy):
(WebKit::WebFramePolicyListenerProxy::didReceiveSafeBrowsingResults):
(WebKit::WebFramePolicyListenerProxy::use):
(WebKit::WebFramePolicyListenerProxy::download):
(WebKit::WebFramePolicyListenerProxy::ignore):

  • UIProcess/WebFramePolicyListenerProxy.h:

(WebKit::WebFramePolicyListenerProxy::create):

  • UIProcess/WebFrameProxy.cpp:

(WebKit::WebFrameProxy::setUpPolicyListenerProxy):

  • UIProcess/WebFrameProxy.h:
  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::decidePolicyForNavigationAction):
(WebKit::WebPageProxy::decidePolicyForNewWindowAction):
(WebKit::WebPageProxy::decidePolicyForResponse):

  • UIProcess/WebPageProxy.h:
  • WebKit.xcodeproj/project.pbxproj:
1:12 PM Changeset in webkit [234512] by timothy@apple.com
  • 7 edits in trunk/Source/WebCore

Text selection color is hard to see in dark mode web views.
​https://bugs.webkit.org/show_bug.cgi?id=188260
rdar://problem/42721294

Reviewed by Simon Fraser.

Stop using blendWithWhite() to transform the AppKit selection color in dark mode.
Using an alpha of 80% gives good contrast, and still works good for selections over images.

  • platform/graphics/Color.cpp:

(WebCore::Color::blendWithWhite const): Mark new colors as semantic if the original is.
(WebCore::Color::colorWithAlpha const): Ditto.

  • rendering/RenderElement.cpp:

(WebCore::RenderElement::selectionBackgroundColor const): Use transformSelectionBackgroundColor.

  • rendering/RenderTheme.cpp:

(WebCore::RenderTheme::activeSelectionBackgroundColor const): Use transformSelectionBackgroundColor.
(WebCore::RenderTheme::inactiveSelectionBackgroundColor const): Ditto.
(WebCore::RenderTheme::transformSelectionBackgroundColor const): Added. Just blend with white.

  • rendering/RenderTheme.h:
  • rendering/RenderThemeMac.h:
  • rendering/RenderThemeMac.mm:

(WebCore::RenderThemeMac::transformSelectionBackgroundColor const): Added. Use an alpha with the color
in dark mode, otherwise fallback to RenderTheme.
(WebCore::RenderThemeMac::systemColor const): Use activeListBoxSelectionBackgroundColor()
and activeSelectionBackgroundColor() instead of caching the colors again. Update hardcoded color.

1:07 PM Changeset in webkit [234511] by Alan Bujtas
  • 2 edits in trunk/Source/WebCore

[LFC][Floating] Use displayBox.rectWithMargin().bottom instead of displayBox.bottom() to where applicable.
​https://bugs.webkit.org/show_bug.cgi?id=188274

Reviewed by Antti Koivisto.

Float placement uses the margin box.

  • layout/FloatingContext.cpp:

(WebCore::Layout::FloatingPair::bottom const):
(WebCore::Layout::Iterator::operator++):
(WebCore::Layout::Iterator::set):

12:56 PM Changeset in webkit [234510] by Alan Bujtas
  • 2 edits in trunk/Source/WebCore

[LFC][Floating] Right aligned float's horizontal candidate position is miscomputed.
​https://bugs.webkit.org/show_bug.cgi?id=188273

Reviewed by Antti Koivisto.

rightAlignedBoxLeft is already shifted with the size of the margin box.

  • layout/FloatingContext.cpp:

(WebCore::Layout::FloatingContext::alignWithFloatings const):

12:47 PM Changeset in webkit [234509] by Alan Bujtas
  • 2 edits in trunk/Source/WebCore

[LFC] Display:Box::rectWithMargin()'s width and height don't include margin left/top.
​https://bugs.webkit.org/show_bug.cgi?id=188272

Reviewed by Antti Koivisto.

  • layout/displaytree/DisplayBox.h:

(WebCore::Display::Box::rectWithMargin const):

12:40 PM Changeset in webkit [234508] by Alan Bujtas
  • 2 edits in trunk/Source/WebCore

[LFC][Floating] Containing block of a float could push the candidate position beyond the current float.
​https://bugs.webkit.org/show_bug.cgi?id=188264

Reviewed by Antti Koivisto.

<div id=first style="float: left; width: 100px; height: 10px;"></div>
<div id=second style="float: left; width: 10px; height: 10px;"></div>
<div style="width: 50px; height: 50px;>

<div id=third style="float: left; width: 10px; height: 10px;"></div>

</div>
In the example above by the time the "third" float comes in, we are already at 110px horizontally. However the containing block of the "third" float forces
the candidate position at 0px. In such cases intersecting won't work since the candidate position is beyond the last(second) float (horizontally).
We could either keep going back to the previous floats on this vertical position and eventually intersect with "first" or just compare the horizontal
positions and ensure that the left/right float is not going to be placed to the left/right of the last float on the current vertical position.

  • layout/FloatingContext.cpp:

(WebCore::Layout::FloatingPair::intersects const):

12:25 PM Changeset in webkit [234507] by rniwa@webkit.org
  • 8 edits in trunk

Implement customElements.upgrade()
​https://bugs.webkit.org/show_bug.cgi?id=183397

Reviewed by Frédéric Wang.

LayoutTests/imported/w3c:

Rebaseline the test now that we're passing.

  • web-platform-tests/custom-elements/custom-element-registry/upgrade-expected.txt:

Source/WebCore:

Added the support to upgrade custom elements directly. Ordinarily, custom elements get upgraded as they are
inserted / connected into a document but some script libraries and authors want to be able to upgrade them before that.
Also see ​https://github.com/w3c/webcomponents/issues/710

Implemented the method as specified at:
​https://html.spec.whatwg.org/multipage/custom-elements.html#dom-customelementregistry-upgrade

When invoked, the upgrade(root) method must run these steps:

  1. Let candidates be a list of all of root's shadow-including inclusive descendant elements, in shadow-including tree order.
  2. For each candidate of candidates, try to upgrade candidate.

Tests: imported/w3c/web-platform-tests/custom-elements/custom-element-registry/upgrade.html

  • dom/CustomElementReactionQueue.cpp:

(WebCore::CustomElementReactionQueue::enqueueElementUpgradeIfDefined): Removed the assertion that the upgraded element
is connected since the whole point of this API is to upgrade a disconnected element.

  • dom/CustomElementRegistry.cpp:

(WebCore::upgradeElementsInShadowIncludingdescendants): Added.
(WebCore::CustomElementRegistry::upgrade): Added.

  • dom/CustomElementRegistry.h: Forward declare DeferredPromise instead of unnecessarily including JSDOMPromiseDeferred.h.
  • dom/CustomElementRegistry.idl:
  • dom/Element.cpp:

(WebCore::Element::insertedIntoAncestor): Moved the assertion here.

12:19 PM Changeset in webkit [234506] by Simon Fraser
  • 9 edits in trunk/Source/WebCore

Convert calc-related enums to 8-bit enum classes
​https://bugs.webkit.org/show_bug.cgi?id=188215

Reviewed by Zalan Bujtas.

This shrinks CSSCalcOperation from 48 to 32 bytes.

  • css/CSSCalculationValue.cpp:

(WebCore::unitCategory):
(WebCore::determineCategory):
(WebCore::resolvedTypeForMinOrMax):
(WebCore::isIntegerResult):
(WebCore::CSSCalcExpressionNodeParser::parseValue):
(WebCore::CSSCalcExpressionNodeParser::parseValueMultiplicativeExpression):
(WebCore::CSSCalcExpressionNodeParser::parseAdditiveValueExpression):
(WebCore::CSSCalcExpressionNodeParser::parseMinMaxExpression):
(WebCore::createBlendHalf):
(WebCore::createCSS):

  • css/CSSCalculationValue.h:
  • css/CSSPrimitiveValue.cpp:

(WebCore::CSSPrimitiveValue::primitiveType const):

  • css/StyleBuilderConverter.h:

(WebCore::StyleBuilderConverter::convertTo100PercentMinusLength):

  • css/parser/CSSPropertyParserHelpers.cpp:

(WebCore::CSSPropertyParserHelpers::CalcParser::consumeNumberRaw):
(WebCore::CSSPropertyParserHelpers::CalcParser::consumePositiveIntegerRaw):
(WebCore::CSSPropertyParserHelpers::consumeInteger):
(WebCore::CSSPropertyParserHelpers::consumePositiveIntegerRaw):
(WebCore::CSSPropertyParserHelpers::consumeNumberRaw):
(WebCore::CSSPropertyParserHelpers::consumeNumber):
(WebCore::CSSPropertyParserHelpers::consumeFontWeightNumber):
(WebCore::CSSPropertyParserHelpers::consumeLength):
(WebCore::CSSPropertyParserHelpers::consumePercent):
(WebCore::CSSPropertyParserHelpers::canConsumeCalcValue):
(WebCore::CSSPropertyParserHelpers::consumeLengthOrPercent):
(WebCore::CSSPropertyParserHelpers::consumeAngle):
(WebCore::CSSPropertyParserHelpers::consumeAngleOrPercent):
(WebCore::CSSPropertyParserHelpers::consumeTime):

  • platform/CalculationValue.cpp:

(WebCore::CalcExpressionNumber::operator== const):
(WebCore::CalcExpressionOperation::evaluate const):
(WebCore::CalcExpressionOperation::operator== const):
(WebCore::CalcExpressionOperation::dump const):
(WebCore::CalcExpressionLength::operator== const):
(WebCore::CalcExpressionBlendLength::CalcExpressionBlendLength):
(WebCore::CalcExpressionBlendLength::operator== const):
(WebCore::operator<<):

  • platform/CalculationValue.h:

(WebCore::CalcExpressionNumber::CalcExpressionNumber):
(WebCore::toCalcExpressionNumber):
(WebCore::CalcExpressionLength::CalcExpressionLength):
(WebCore::toCalcExpressionLength):
(WebCore::CalcExpressionOperation::CalcExpressionOperation):
(WebCore::toCalcExpressionOperation):
(WebCore::toCalcExpressionBlendLength):

  • platform/Length.cpp:

(WebCore::convertTo100PercentMinusLength):

12:06 PM Changeset in webkit [234505] by mitz@apple.com
  • 3 edits in trunk/Tools

Fixed WebKit.AttrStyle timing out on build.webkit.org.

  • TestWebKitAPI/Tests/WebKitCocoa/AttrStyle.html: Changed from a URL that requires network access to one that doesn’t.
  • TestWebKitAPI/Tests/WebKitCocoa/AttrStyle.mm:

(TEST): Updated expected value.

11:32 AM Changeset in webkit [234504] by Wenson Hsieh
  • 7 edits
    3 adds in trunk

[iOS] Keyboard becomes unresponsive after pressing delete while pressing down on a character key with accents
​https://bugs.webkit.org/show_bug.cgi?id=188251
<rdar://problem/37842108>

Reviewed by Tim Horton.

Source/WebKit:

Fixes a bug in key event handling where invoking -handleKeyWebEvent:withCompletionHandler: from within the
completion callback of a previous call to -handleKeyWebEvent:withCompletionHandler: would cause the completion
callback to be cleared out prematurely. In some cases (as described in the title of this bug), UIKit exercises
this codepath and subsequently hangs due to their completion block never getting invoked by WebKit.

Test: KeyboardInputTests.CanHandleKeyEventInCompletionHandler

  • UIProcess/ios/WKContentViewInteraction.h:
  • UIProcess/ios/WKContentViewInteraction.mm:

(-[WKContentView handleKeyWebEvent:withCompletionHandler:]):
(-[WKContentView _didHandleKeyEvent:eventWasHandled:]):

Tools:

Adds a new API test to verify that clients (in this case, UIKit) is allowed to invoke
-handleKeyWebEvent:withCompletionHandler: within the completion block of a prior invocation.

  • TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
  • TestWebKitAPI/Tests/ios/KeyboardInputTestsIOS.mm: Added.

(TestWebKitAPI::TEST):

  • TestWebKitAPI/Tests/ios/TestInputDelegate.h: Added.

Pull some logic used to force an input session to start out from an existing API test file
(WKWebViewAutofillTests) and into a separate helper class that is used by both the existing API tests and the
new keyboard input test.

  • TestWebKitAPI/Tests/ios/WKWebViewAutofillTests.mm:

(-[AutofillTestView initWithFrame:]):
(TestWebKitAPI::TEST):
(-[TestInputDelegate init]): Deleted.
(-[TestInputDelegate _webView:focusShouldStartInputSession:]): Deleted.

  • TestWebKitAPI/ios/UIKitSPI.h:

Add some UIKit SPI utilized by the new API test.

10:43 AM Changeset in webkit [234503] by cturner@igalia.com
  • 2 edits in trunk/Source/WebCore

Handle zero-sized ISOMP4 boxes appropriately
​https://bugs.webkit.org/show_bug.cgi?id=188256

Reviewed by Jer Noble.

According to ISO/IEC 14496-12:2012(E), when the Box classes' size
field is zero, the implied size of the box extends to the end of
the file. If this detail is not taken into account, CENC
sanitization can incorrectly report an invalid box size, since 0
!= the number of bytes in this box, specifically, the data layout
of Box is as follows,

aligned(8) class Box (unsigned int(32) boxtype,

optional unsigned int(8)[16] extended_type) {

unsigned int(32) size;
unsigned int(32) type = boxtype;
if (size==1) {

unsigned int(64) largesize;

} else if (size==0) { This is the case now handled.

box extends to end of file

}
if (boxtype==‘uuid’) {

unsigned int(8)[16] usertype = extended_type;

}

}

Tested by imported/w3c/web-platform-tests/encrypted-media/clearkey-generate-request-disallowed-input.https.html

  • platform/graphics/iso/ISOBox.cpp:

(WebCore::ISOBox::peekBox): Check if the parsed size is zero, and
if it is, the size is calculated as the total number of bytes in
the incoming DataView.
(WebCore::ISOBox::parse): Ditto.

9:42 AM Changeset in webkit [234502] by mitz@apple.com
  • 15 edits
    2 adds in trunk

Optionally expose Attr::style to JavaScript
​https://bugs.webkit.org/show_bug.cgi?id=188226
<rdar://problem/42818113>

Reviewed by Darin Adler.

Source/WebCore:

Test: TestWebKitAPI/Tests/WebKitCocoa/AttrStyle.mm

  • dom/Attr.cpp: Update comment.
  • dom/Attr.idl: Define the style attribute, enabled at runtime by the AttrStyle feature.
  • page/RuntimeEnabledFeatures.h:

(WebCore::RuntimeEnabledFeatures::setAttrStyleEnabled): Added this accessor.
(WebCore::RuntimeEnabledFeatures::attrStyleEnabled const): Ditto.

Source/WebKit:

  • Shared/WebProcessCreationParameters.cpp:

(WebKit::WebProcessCreationParameters::encode const): Encode new attrStyleEnabled member.
(WebKit::WebProcessCreationParameters::decode): Decode new attrStyleEnabled member.

  • Shared/WebProcessCreationParameters.h: Define new attrStyleEnabled member, initialized to false.
  • UIProcess/API/APIProcessPoolConfiguration.cpp:

(API::ProcessPoolConfiguration::copy): Copy new m_attrStyleEnabled member.

  • UIProcess/API/APIProcessPoolConfiguration.h: Define new m_attrStyleEnabled member.
  • UIProcess/API/Cocoa/_WKProcessPoolConfiguration.h: Declare new attrStyleEnabled boolean property.
  • UIProcess/API/Cocoa/_WKProcessPoolConfiguration.mm:

(-[_WKProcessPoolConfiguration attrStyleEnabled]): New accessor.
(-[_WKProcessPoolConfiguration setAttrStyleEnabled:]): Ditto.

  • UIProcess/WebProcessPool.cpp:

(WebKit::WebProcessPool::initializeNewWebProcess): Initialize parameters.attrStyleEnabled

from the configuration.

  • WebProcess/WebProcess.cpp:

(WebKit::WebProcess::initializeWebProcess): Set the attrStyleEnabled runtime feature based

on the creation parameters.

Tools:

  • TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
  • TestWebKitAPI/Tests/WebKitCocoa/AttrStyle.html: Added.
  • TestWebKitAPI/Tests/WebKitCocoa/AttrStyle.mm: Added.

(TEST):

8:45 AM Changeset in webkit [234501] by david_fenton@apple.com
  • 116 edits in trunk

Unreviewed, rolling out r234489.

Caused 50+ crashes and 60+ API failures on iOS

Reverted changeset:

"[WTF] Rename String::format to String::deprecatedFormat"
​https://bugs.webkit.org/show_bug.cgi?id=188191
​https://trac.webkit.org/changeset/234489

8:01 AM Changeset in webkit [234500] by Alan Bujtas
  • 4 edits in trunk/Source/WebCore

[LFC][Floating] Do not pass formatting root to FloatingContext
​https://bugs.webkit.org/show_bug.cgi?id=188257

Reviewed by Antti Koivisto.

If we ever need it, we can get it from FloatingState.

  • layout/FloatingContext.cpp:

(WebCore::Layout::FloatingContext::FloatingContext):

  • layout/FloatingContext.h:
  • layout/blockformatting/BlockFormattingContext.cpp:

(WebCore::Layout::BlockFormattingContext::layout const):

7:28 AM Changeset in webkit [234499] by Alan Bujtas
  • 6 edits in trunk/Source/WebCore

[LFC][Floating] Convert all coordinates relative to the formatting context root.
​https://bugs.webkit.org/show_bug.cgi?id=188249

Reviewed by Antti Koivisto.

The idea here is that all the float related computation should be happening in the same coordinate system.
FloatItem holds the transformed Display::Box for both the layout box and its containing block.

  • layout/FloatingContext.cpp:

(WebCore::Layout::begin):
(WebCore::Layout::end):
(WebCore::Layout::FloatingContext::computePosition const):
(WebCore::Layout::FloatingContext::floatingPosition const):
(WebCore::Layout::FloatingContext::initialVerticalPosition const):
(WebCore::Layout::FloatingContext::alignWithContainingBlock const):
(WebCore::Layout::FloatingContext::alignWithFloatings const):
(WebCore::Layout::FloatingContext::toContainingBlock const):
(WebCore::Layout::FloatingPair::FloatingPair):
(WebCore::Layout::FloatingPair::left const):
(WebCore::Layout::FloatingPair::right const):
(WebCore::Layout::Iterator::Iterator):
(WebCore::Layout::previousFloatingIndex):
(WebCore::Layout::Iterator::operator++):
(WebCore::Layout::Iterator::set):

  • layout/FloatingContext.h:
  • layout/FloatingState.cpp:

(WebCore::Layout::FloatingState::FloatItem::FloatItem):
(WebCore::Layout::FloatingState::remove):
(WebCore::Layout::FloatingState::append):

  • layout/FloatingState.h:

(WebCore::Layout::FloatingState::isEmpty const):
(WebCore::Layout::FloatingState::FloatItem::layoutBox const):
(WebCore::Layout::FloatingState::FloatItem::containingBlock const):
(WebCore::Layout::FloatingState::FloatItem::displayBox const):
(WebCore::Layout::FloatingState::FloatItem::containingBlockDisplayBox const):
(WebCore::Layout::FloatingState::floats const):
(WebCore::Layout::FloatingState::last const):
(WebCore::Layout::FloatingState::root const):
(WebCore::Layout::FloatingState::floatings const): Deleted.

  • layout/displaytree/DisplayBox.h:
5:21 AM Changeset in webkit [234498] by cturner@igalia.com
  • 2 edits in trunk/Source/WebCore

[GStreamer] Dynamically adjust blocksize
​https://bugs.webkit.org/show_bug.cgi?id=188194

Reviewed by Xabier Rodriguez-Calvar.

Update the blocksize depending on how much is obtained from a read
of the input stream. This avoids doing too many reads in small
chunks when larger amounts of data are available and also prevents
using a very large memory area to read a small chunk of data.

Adapted from ​https://bugzilla.gnome.org/show_bug.cgi?id=767833

No tests new tests since no change in functionality.

  • platform/graphics/gstreamer/WebKitWebSourceGStreamer.cpp:

(webkit_web_src_init):
(CachedResourceStreamingClient::checkUpdateBlocksize):
(CachedResourceStreamingClient::dataReceived):

2:23 AM Changeset in webkit [234497] by cturner@igalia.com
  • 2 edits in trunk/Source/WebCore

[GStreamer] Stop pushing buffers when seeking status changes
​https://bugs.webkit.org/show_bug.cgi?id=188193

Reviewed by Xabier Rodriguez-Calvar.

After switching to splitting buffers into smaller block sizes in

​https://bugs.webkit.org/show_bug.cgi?id=182829

It was found that during the individual buffer pushes, the seeking
status could change behind our backs from another thread. When
this happens, buffers from incorrect offsets would find their way
into appsrc and eventually the demuxer itself, which would start
parsing from a random place and at best give a confusing error
message.

The solution here is break from pushing buffers when the seeking
status has been has changed. Flushes will clear out what we've
already delivered into the appsrc, then we must make sure to not
continue sending buffers in there after the flush.

No new tests since this is a timing bug.

  • platform/graphics/gstreamer/WebKitWebSourceGStreamer.cpp:

(CachedResourceStreamingClient::dataReceived):

Note: See TracTimeline for information about the timeline view.