⚠ Archived content — this site is no longer maintained.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

Timeline



Nov 19, 2018:

11:09 PM Changeset in webkit [238391] by mark.lam@apple.com
  • 5 edits
    1 add in trunk

globalFuncImportModule() should return a promise when it clears exceptions.
​https://bugs.webkit.org/show_bug.cgi?id=191792
<rdar://problem/46090763>

Reviewed by Michael Saboff.

JSTests:

  • stress/global-import-function-should-return-a-promise-when-clearing-exceptions.js: Added.

Source/JavaScriptCore:

If we're clearing the exceptions in a CatchScope, then it means that we've handled
the exception, and is able to proceed in a normal manner. Hence, we should not
return the empty JSValue in this case: instead, we should return a Promise as
expected by import's API.

The only time when we can't return a promise is when we fail to create a Promise.
In that case, we should be propagating the exception.

Hence, globalFuncImportModule() contains a ThrowScope (for propagating the
exception that arises from failure to create the Promise) wrapping a CatchScope
(for catching any exception that arises from failure to execute the import).

Also fixed similar issues, and some exception check issues in JSModuleLoader and
the jsc shell.

  • jsc.cpp:

(GlobalObject::moduleLoaderImportModule):
(GlobalObject::moduleLoaderFetch):

  • runtime/JSGlobalObjectFunctions.cpp:

(JSC::globalFuncImportModule):

  • runtime/JSModuleLoader.cpp:

(JSC::JSModuleLoader::loadAndEvaluateModule):
(JSC::JSModuleLoader::loadModule):
(JSC::JSModuleLoader::requestImportModule):
(JSC::JSModuleLoader::importModule):
(JSC::JSModuleLoader::resolve):
(JSC::JSModuleLoader::fetch):
(JSC::moduleLoaderParseModule):
(JSC::moduleLoaderResolveSync):

10:46 PM Changeset in webkit [238390] by Fujii Hironori
  • 2 edits in trunk/Source/WebCore

REGRESSION(r238350) [curl] CertificateInfo.h: error: template specialization requires 'template<>'
​https://bugs.webkit.org/show_bug.cgi?id=191849

Unreviewed build fix for clang-cl builds.

No new tests because there's no behaviour change.

  • platform/network/curl/CertificateInfo.h:

(WTF::Persistence::Coder<WebCore::CertificateInfo>::encode):
(WTF::Persistence::Coder<WebCore::CertificateInfo>::decode):

7:04 PM Changeset in webkit [238389] by don.olmstead@sony.com
  • 2 edits in trunk/Source/WebKit

Simplify platform check in WebEditorClient
​https://bugs.webkit.org/show_bug.cgi?id=191848

Reviewed by Fujii Hironori.

!PLATFORM(GTK) && !PLATFORM(COCOA) && !PLATFORM(WPE) can be better
represented as PLATFORM(WIN).

  • WebProcess/WebCoreSupport/WebEditorClient.cpp:
6:16 PM Changeset in webkit [238388] by achristensen@apple.com
  • 23 edits
    1 add in trunk

Add SPI to disable JIT in a WKWebView
​https://bugs.webkit.org/show_bug.cgi?id=191822
<rdar://problem/28119360>

Reviewed by Geoffrey Garen.

Source/JavaScriptCore:

  • jit/ExecutableAllocator.cpp:

(JSC::jitDisabled):
(JSC::allowJIT):
(JSC::ExecutableAllocator::setJITEnabled):

  • jit/ExecutableAllocator.h:

(JSC::ExecutableAllocator::setJITEnabled):

Source/WebKit:

  • Shared/EntryPointUtilities/mac/XPCService/XPCServiceEntryPoint.h:

(WebKit::XPCServiceInitializer):

  • UIProcess/API/APIProcessPoolConfiguration.h:
  • UIProcess/API/Cocoa/WKWebView.mm:

(-[WKWebView _canUseJIT:]):

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

(-[_WKProcessPoolConfiguration enableJIT]):
(-[_WKProcessPoolConfiguration setEnableJIT:]):

  • UIProcess/Launcher/ProcessLauncher.h:

(WebKit::ProcessLauncher::Client::enableJIT const):

  • UIProcess/Launcher/mac/ProcessLauncherMac.mm:

(WebKit::ProcessLauncher::launchProcess):

  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::canUseJIT):

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

(WebKit::WebProcessProxy::enableJIT const):

  • UIProcess/WebProcessProxy.h:

(WebKit::WebProcessProxy::processPool const):
(WebKit::WebProcessProxy::processPool): Deleted.

  • WebProcess/WebProcess.cpp:

(WebKit::WebProcess::canUseJIT):

  • WebProcess/WebProcess.h:
  • WebProcess/WebProcess.messages.in:

Tools:

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

(TEST):

6:02 PM Changeset in webkit [238387] by basuke.suzuki@sony.com
  • 9 edits
    1 copy
    4 adds in trunk

[Curl] Add API for CertificateInfo.
​https://bugs.webkit.org/show_bug.cgi?id=191647

Reviewed by Alex Christensen.

Source/WebCore:

Minor changes for WebKit API.

Tests: TestWebKitAPI/Tests/WebKit/curl/Certificates.cpp

  • platform/network/curl/CertificateInfo.h:
  • platform/network/curl/CertificateInfoCurl.cpp:

(WebCore::CertificateInfo::CertificateInfo):
(WebCore::CertificateInfo::makeCertificate):

  • platform/network/curl/CurlSSLVerifier.cpp:

(WebCore::BIOHolder::asCertificate):

Source/WebKit:

Added API for CertificateInfo to create, fetch contents.

Tests: TestWebKitAPI/Tests/WebKit/curl/Certificates.cpp

  • PlatformWin.cmake:
  • Shared/API/c/curl/WKCertificateInfoCurl.cpp: Added.

(WKCertificateInfoCreateWithCertficateChain):
(WKCertificateInfoGetVerificationError):
(WKCertificateInfoGetCertificateChainSize):
(WKCertificateInfoCopyCertificateAtIndex):

  • Shared/API/c/curl/WKCertificateInfoCurl.h: Copied from Source/WebCore/platform/network/curl/CertificateInfoCurl.cpp.
  • Shared/curl/WebCoreArgumentCodersCurl.cpp:

(IPC::ArgumentCoder<CertificateInfo>::encode):
(IPC::ArgumentCoder<CertificateInfo>::decode):

Tools:

Added test for new WKCertificateInfo API.

  • TestWebKitAPI/PlatformWin.cmake:
  • TestWebKitAPI/Tests/WebKit/curl/Certificates.cpp: Added.

(TestWebKitAPI::Curl::makeCertificateInfo):
(TestWebKitAPI::Curl::isSamePEM):
(TestWebKitAPI::Curl::TEST):

5:37 PM Changeset in webkit [238386] by Fujii Hironori
  • 2 edits in trunk/Source/JavaScriptCore

[MSVC] X86Assembler.h(108): error C2666: 'WebCore::operator -': 7 overloads have similar conversions
​https://bugs.webkit.org/show_bug.cgi?id=189467
<rdar://problem/44290945>

Reviewed by Mark Lam.

This issue has happened several times. And, it seems that it will
take more time for Microsoft to fix the MSVC bug. We need a
effective workaround not to repeat this issue until they fix MSVC.

Remove ": int8_t" of RegisterID only for COMPILER(MSVC).

  • assembler/X86Assembler.h: Added JSC_X86_ASM_REGISTER_ID_ENUM_BASE_TYPE macro.
5:15 PM Changeset in webkit [238385] by Devin Rousso
  • 2 edits in trunk/Source/WebInspectorUI

Web Inspector: Audit: removing and re-adding the tab will duplicate tests
​https://bugs.webkit.org/show_bug.cgi?id=191843

Reviewed by Matt Baker.

  • UserInterface/Controllers/AuditManager.js:

(WI.AuditManager.prototype.loadStoredTests):
If we already have tests, don't attempt to load any stored tests.

5:11 PM Changeset in webkit [238384] by don.olmstead@sony.com
  • 11 edits
    6 copies
    1 add
    1 delete in trunk/Source/WebKit

Add USE(LIBWPE) for WebKit events
​https://bugs.webkit.org/show_bug.cgi?id=191842

Reviewed by Michael Catanzaro.

Moves all WPE events to USE(LIBWPE) instead of PLATFORM(WPE). Updates
filenames to LibWPE.

  • PlatformWPE.cmake:
  • Shared/NativeWebKeyboardEvent.h:
  • Shared/NativeWebMouseEvent.h:
  • Shared/NativeWebTouchEvent.h:
  • Shared/NativeWebWheelEvent.h:
  • Shared/WebCoreArgumentCoders.cpp:
  • Shared/WebCoreArgumentCoders.h:
  • Shared/WebEvent.h:
  • Shared/WebKeyboardEvent.cpp:

(WebKit::WebKeyboardEvent::WebKeyboardEvent):

  • Shared/libwpe/NativeWebKeyboardEventLibWPE.cpp: Renamed from Source/WebKit/Shared/wpe/NativeWebKeyboardEventWPE.cpp.
  • Shared/libwpe/NativeWebMouseEventLibWPE.cpp: Renamed from Source/WebKit/Shared/wpe/NativeWebMouseEventWPE.cpp.
  • Shared/libwpe/NativeWebTouchEventLibWPE.cpp: Renamed from Source/WebKit/Shared/wpe/NativeWebTouchEventWPE.cpp.
  • Shared/libwpe/NativeWebWheelEventLibWPE.cpp: Renamed from Source/WebKit/Shared/wpe/NativeWebWheelEventWPE.cpp.
  • Shared/libwpe/WebEventFactory.cpp: Renamed from Source/WebKit/Shared/wpe/WebEventFactory.cpp.
  • Shared/libwpe/WebEventFactory.h: Renamed from Source/WebKit/Shared/wpe/WebEventFactory.h.
  • SourcesWPE.txt:
4:26 PM Changeset in webkit [238383] by Wenson Hsieh
  • 2 edits in trunk/Tools

Unreviewed, fix the internal iOS debug build after r238380

UICalloutBar should only be declared when the public (non-internal) SDK is used to build WebKit; otherwise, we
end up with duplicate declarations of UICalloutBar.

  • TestWebKitAPI/ios/UIKitSPI.h:
3:33 PM Changeset in webkit [238382] by dino@apple.com
  • 20 edits
    11 copies
    3 adds
    1 delete in trunk

[WebGPU] Begin implementing WebGPUTexture, WebGPUTextureView, and WebGPUTextureFormatEnum, and WebGPUSwapChain::configure upgrades
​https://bugs.webkit.org/show_bug.cgi?id=191794

Patch by Justin Fan <Justin Fan> on 2018-11-19
Reviewed by Dean Jackson.

Source/WebCore:

Test: webgpu/textures-textureviews.html

Implement basic functionality for getting the next WebGPUTexture and TextureView from the WebGPURenderingContext
to use as a render destination for the next draw call. Also introduce WebGPUTextureFormatEnum and the ability to
configure the context with a chosen texture format.

  • CMakeLists.txt:
  • DerivedSources.make:
  • Modules/webgpu/GPUSwapChain.h: Texture/Pixel format can now be set.
  • Modules/webgpu/GPUTexture.h: Added. Interface to a MTLTexture.
  • Modules/webgpu/GPUTextureFormatEnum.h: Added.
  • Modules/webgpu/WebGPUDevice.cpp:

(WebCore::WebGPUDevice::createRenderPipeline const): Removed now-unnecessary enum class cast.

  • Modules/webgpu/WebGPURenderPipelineDescriptor.h: Rather than duplicate GPURenderPipelineDescriptor::PrimitiveTopology, alias to it.
  • Modules/webgpu/WebGPUSwapChain.cpp:

(WebCore::WebGPUSwapChain::configure): Can now specify a specific texture format for the underlying CAMetalLayer.
(WebCore::WebGPUSwapChain::getNextTexture): Added. Request the next drawable texture.

  • Modules/webgpu/WebGPUSwapChain.h: Expose getNextTexture().
  • Modules/webgpu/WebGPUSwapChain.idl:
  • Modules/webgpu/WebGPUTexture.cpp: Added.

(WebCore::WebGPUTexture::create):
(WebCore::WebGPUTexture::WebGPUTexture):
(WebCore::WebGPUTexture::createDefaultTextureView):

  • Modules/webgpu/WebGPUTexture.h: Added.
  • Modules/webgpu/WebGPUTexture.idl: Added.
  • Modules/webgpu/WebGPUTextureFormatEnum.h: Added. Type alias for GPUTextureFormatEnum.
  • Modules/webgpu/WebGPUTextureFormatEnum.idl: Added. Used to represent any texture format used by WebGPU.
  • Modules/webgpu/WebGPUTextureView.cpp: Added.

(WebCore::WebGPUTextureView::create):
(WebCore::WebGPUTextureView::WebGPUTextureView):

  • Modules/webgpu/WebGPUTextureView.h: Added.
  • Modules/webgpu/WebGPUTextureView.idl: Added.
  • Modules/webgpu/cocoa/GPUSwapChainMetal.mm:

(WebCore::GPUSwapChain::create):
(WebCore::GPUSwapChain::setFormat): Called by WebGPUSwapChain::configure().
(WebCore::GPUSwapChain::getNextTexture):

  • Modules/webgpu/cocoa/GPUTextureFormatEnumMetal.mm: Added.

(WebCore::convertAndValidate): Convert the WebGPUTextureFormatEnum to a MTLPixelFormat.

  • Modules/webgpu/cocoa/GPUTextureMetal.mm: Added.

(WebCore::GPUTexture::create):
(WebCore::GPUTexture::GPUTexture):
(WebCore::GPUTexture::createDefaultTextureView): Uses the pixelFormat of the original texture.

  • Sources.txt:
  • SourcesCocoa.txt:
  • WebCore.xcodeproj/project.pbxproj:
  • bindings/js/WebCoreBuiltinNames.h:

LayoutTests:

Did some refactoring to better share code among tests, and added basic functionality test for
WebGPUTextures and WebGPUTextureView, to be fleshed out when their descriptor objects are implemented.

  • webgpu/command-buffers.html: Refactored to use basic-webgpu-functions.js.
  • webgpu/js/basic-webgpu-functions.js: Added.

(vertex.Vertex.vertex_main): The basic MSL code has been moved into this file.
(fragment.float4.fragment_main):
(runWebGPUTests): Creates a WebGPURenderingContext on a default canvas and runs its provided functions.
(async.runWebGPUTestsOnCanvas):
(async.setUpContexts): Tests proper creation of a basic WebGPURenderingContext.
(setUpModule): Creates a basic WebGPUShaderModule.
(setUpPipelineDescriptor): Creates a basic WebGPURenderPipelineDescriptor.
(setUpPipeline): Creates a basic WebGPURenderPipeline.
(render):

  • webgpu/js/set-up-webgpu-contexts.js: Moved into basic-webgpu-functions.js.
  • webgpu/render-pipelines-expected.txt:
  • webgpu/render-pipelines.html: Refactored to use basic-webgpu-functions.js.
  • webgpu/shader-modules.html: Refactored to use basic-webgpu-functions.js.
  • webgpu/textures-textureviews-expected.txt: Added.
  • webgpu/textures-textureviews.html: Added.
  • webgpu/webgpu-basics.html: Refactored to use basic-webgpu-functions.js.
2:52 PM Changeset in webkit [238381] by david_quesada@apple.com
  • 6 edits in trunk

EXC_BAD_ACCESS when invoking a DownloadProxy's destination decision handler after the download has been canceled
​https://bugs.webkit.org/show_bug.cgi?id=191762
rdar://problem/46151509

Reviewed by Dean Jackson.

Source/WebKit:

When the DownloadClient calls the decision handler with a destination path, check if
m_processPool is null before trying to access its network process. This can happen
if a download is canceled before the client decides its destination.

  • UIProcess/Downloads/DownloadProxy.cpp:

(WebKit::DownloadProxy::decideDestinationWithSuggestedFilenameAsync):

Tools:

  • TestWebKitAPI/Tests/WebKitCocoa/Download.mm:

Enable the Download API test on iOS, since the platform supports downloads. All the
tests pass already, except for one which was written using AppKit-specific code:

TEST(_WKDownload, RedirectedDownload):

Use a more platform-agnostic approach to starting the download in this API test.
Instead of manually triggering an NSMenu item to download a file from a link, the
test will simulate a user-initiated click on the link, and the navigation delegate
will direct the web view to start a download based on the link's navigation action.
Additionally, remove the manual creation of a new NSWindow. TestWKWebView makes its
own UI/NSWindow.

TEST(_WKDownload, DownloadCanceledWhileDecidingDestination):

Add an API test _WKDownload.DownloadCanceledWhileDecidingDestination to simulate the
conditions that would trigger this crash - handling a download's -decideDestination…
delegate call by canceling the download, waiting for the UI process to be notified
that the download has been canceled, and calling the decision handler. This should
not crash.

(-[CancelDownloadWhileDecidingDestinationDelegate _downloadDidFinish:]):
(-[CancelDownloadWhileDecidingDestinationDelegate _download:didFailWithError:]):
(-[CancelDownloadWhileDecidingDestinationDelegate _downloadDidCancel:]):
(-[CancelDownloadWhileDecidingDestinationDelegate _download:decideDestinationWithSuggestedFilename:completionHandler:]):
(-[UIDownloadAsFileTestDelegate _webView:contextMenu:forElement:]): Deleted.

  • TestWebKitAPI/cocoa/TestWKWebView.h:
  • TestWebKitAPI/cocoa/TestWKWebView.mm:

(-[TestWKWebView objectByEvaluatingJavaScriptWithUserGesture:]):

Add a user-initated version of -objectByEvaluatingJavaScript:. This is needed in
order to maintain the behavior of the RedirectedDownload test, which verifies the
state of the _WKDownload.wasUserInitiated property.

1:37 PM Changeset in webkit [238380] by Wenson Hsieh
  • 3 edits in trunk/Tools

[iOS Debug] Assertion failure in NSComparisonResult UIContentSizeCategoryCompareToCategory()
​https://bugs.webkit.org/show_bug.cgi?id=190401
<rdar://problem/45195285>

Reviewed by Dean Jackson.

Work around sporadic UIKit assertions when running API tests that may present a UICalloutBar. Currently, UIKit
asks for the UIApp’s -preferredContentSizeCategory when showing or laying out the callout bar. However,
TestWebKitAPI lacks a UIApp altogether, and forcing one to be initialized via UIApplicationInstantiateSingleton
causes numerous other UIKit and WebKit assertions to be hit.

In lieu of TestWebKitAPI becoming a real UI application (see ​https://webkit.org/b/175204), simply don't allow
the callout bar to be initialized in TestWKWebView-based API tests. This encompasses all API tests which
have some chance of causing the callout bar to be shown by making a ranged text selections during the test.

  • TestWebKitAPI/cocoa/TestWKWebView.mm:

(suppressUICalloutBar):
(-[TestWKWebView initWithFrame:configuration:addToWindow:]):

  • TestWebKitAPI/ios/UIKitSPI.h:
1:22 PM Changeset in webkit [238379] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit

[cairo] BackingStore::incorporateUpdate should use CompositeCopy to support drawsBackground=false in general Cairo ports
​https://bugs.webkit.org/show_bug.cgi?id=191577

Patch by Tomoki Imai <Tomoki Imai> on 2018-11-19
Reviewed by Carlos Garcia Campos.

We should use CompositeCopy operator to update bitmap to overwrite existing bitmap even if source contains
transparent parts. It enables ports which uses Cairo to set drawsBackground to false.
GTK ports has custom background extension, and in such case, we need to use CompositeSourceOver to retain
custom background.

  • UIProcess/cairo/BackingStoreCairo.cpp:

(WebKit::BackingStore::incorporateUpdate): Use CompositeCopy operator to update bitmap.

12:51 PM Changeset in webkit [238378] by Joseph Pecoraro
  • 30 edits in trunk/Source

Web Inspector: "Reload Web Inspector" button no longer partially works
​https://bugs.webkit.org/show_bug.cgi?id=191773
<rdar://problem/46139932>

Reviewed by Devin Rousso.

Source/WebCore:

  • inspector/InspectorFrontendClient.h:
  • inspector/InspectorFrontendHost.cpp:

(WebCore::InspectorFrontendHost::reopen):

  • inspector/InspectorFrontendHost.h:
  • inspector/InspectorFrontendHost.idl:

Provide a host call to reopen an inspector window to reload it.

  • testing/Internals.cpp:

Stub implementation, this is not used in tests.

Source/WebInspectorUI:

  • UserInterface/Base/Main.js:
  • UserInterface/Debug/UncaughtExceptionReporter.js:

(sheetElement.innerHTML):

  • UserInterface/Views/SettingsTabContentView.js:

Source/WebKit:

  • UIProcess/RemoteWebInspectorProxy.cpp:

(WebKit::RemoteWebInspectorProxy::load):
(WebKit::RemoteWebInspectorProxy::reopen):

  • UIProcess/RemoteWebInspectorProxy.h:
  • UIProcess/RemoteWebInspectorProxy.messages.in:
  • WebProcess/WebPage/RemoteWebInspectorUI.cpp:

(WebKit::RemoteWebInspectorUI::reopen):

  • WebProcess/WebPage/RemoteWebInspectorUI.h:
  • WebProcess/WebPage/WebInspector.cpp:

(WebKit::WebInspector::reopen):

  • WebProcess/WebPage/WebInspector.h:
  • WebProcess/WebPage/WebInspector.messages.in:
  • WebProcess/WebPage/WebInspectorUI.cpp:

(WebKit::WebInspectorUI::reopen):

  • WebProcess/WebPage/WebInspectorUI.h:

Source/WebKitLegacy/ios:

  • WebCoreSupport/WebInspectorClientIOS.mm:

(WebInspectorFrontendClient::reopen):

Source/WebKitLegacy/mac:

  • WebCoreSupport/WebInspectorClient.h:
  • WebCoreSupport/WebInspectorClient.mm:

(WebInspectorFrontendClient::reopen):

Source/WebKitLegacy/win:

  • WebCoreSupport/WebInspectorClient.cpp:

(WebInspectorFrontendClient::reopen):

  • WebCoreSupport/WebInspectorClient.h:
12:42 PM Changeset in webkit [238377] by commit-queue@webkit.org
  • 3 edits
    2 adds in trunk

Setting document.title should have no effect for non SVG/HTML documents
​https://bugs.webkit.org/show_bug.cgi?id=191643

Patch by Rob Buis <​rbuis@igalia.com> on 2018-11-19
Reviewed by Chris Dumez.

LayoutTests/imported/w3c:

  • web-platform-tests/html/dom/documents/dom-tree-accessors/document.title-not-in-html-svg-expected.txt: Added.
  • web-platform-tests/html/dom/documents/dom-tree-accessors/document.title-not-in-html-svg.html: Added.

Source/WebCore:

Setting document.title should have no effect for non SVG/HTML documents,
see ​https://html.spec.whatwg.org/multipage/dom.html#document.title.

Behavior matches Firefox and Chrome.

Test: imported/w3c/web-platform-tests/html/dom/documents/dom-tree-accessors/document.title-not-in-html-svg.html

  • dom/Document.cpp:

(WebCore::Document::setTitle):

8:49 AM Changeset in webkit [238376] by yusukesuzuki@slowstart.org
  • 5 edits in trunk/Source/JavaScriptCore

[WebAssembly] I64 arguments / return value check should be moved from callWebAssemblyFunction to JSToWasm wrapper
​https://bugs.webkit.org/show_bug.cgi?id=190512

Reviewed by Keith Miller.

This patch moves I64 arguments / return value check from callWebAssemblyFunction to JSToWasm wrapper. Since this
check can be done when compiling the function, we should encode the result into the generated wrapper instead of
checking every time we call callWebAssemblyFunction. This change is also one of the steps removing callWebAssemblyFunction
entirely.

  • wasm/WasmExceptionType.h:
  • wasm/js/JSToWasm.cpp:

(JSC::Wasm::createJSToWasmWrapper):

  • wasm/js/WebAssemblyFunction.cpp:

(JSC::callWebAssemblyFunction):

  • wasm/js/WebAssemblyWrapperFunction.cpp:

(JSC::callWebAssemblyWrapperFunction):

8:31 AM Changeset in webkit [238375] by Wenson Hsieh
  • 7 edits
    2 adds in trunk

Dragging image with a border-image larger than the image element crashes
​https://bugs.webkit.org/show_bug.cgi?id=191817
<rdar://problem/46159222>

Reviewed by Ryosuke Niwa.

Source/WebCore:

When dragging an image element, if the image element has:

(1) box-sizing: border-box;
(2) a border-image
(3) a border-top-width that is at least as large as the height of the element and/or a border-left-width that is

at least as large as the width of the element

...then upon drag, we will fail to create a suitable drag image using the bounding box of the image element
since the size is empty, thereby causing a crash. To fix this, we bail out of this bounding-rect-dependent
codepath for generating a drag image in the case where the bounding rect is empty, and instead fall back to an
icon representation for the drag image.

Test: fast/events/drag-image-with-border-image.html

  • page/DragController.cpp:

(WebCore::DragController::doImageDrag):

LayoutTests:

Verifies that an image that meets the pathological criteria described in Source/WebCore/ChangeLog can still be
dragged and dropped into an editable area.

  • fast/events/drag-image-with-border-image.html: Added.
  • platform/gtk/TestExpectations:
  • platform/ios/TestExpectations:
  • platform/mac-wk2/TestExpectations:
  • platform/wpe/TestExpectations:

Enable this test only in WebKit1.

8:26 AM Changeset in webkit [238374] by Fujii Hironori
  • 3 edits in trunk/Tools

REGRESSION(r238016)[GTK][TestWebKitAPI][Ninja] TestJSC can't compile "Bytecodes.h: No such file or directory"
​https://bugs.webkit.org/show_bug.cgi?id=191692

Reviewed by Michael Catanzaro.

r238016 added a new #include, and requires JSC's derived headers
to be copied into the forwarding header directory.

WEBKIT_MAKE_FORWARDING_HEADERS is using POST_BUILD event to copy
derived headers to the forwarding header directory. This is done
after linking JSC library.

GTK port has TestJSC which has a dependency to JSC static library.
And, this dependency is a link time dependency which means
compiling source files of TestJSC starts before the POST_BUILD
event of JSC. Thus, required header can't be found.

Add an intermediate target between TestJSC and JavaScriptCore to
ensure derived headers are copied into the forwarding header
directory.

  • TestWebKitAPI/PlatformGTK.cmake: Added a new custom target pre-TestJSC.
  • TestWebKitAPI/PlatformWPE.cmake: Ditto.
6:00 AM Changeset in webkit [238373] by guijemont@igalia.com
  • 4 edits in trunk/JSTests

Skip new memory-hungry tests on memory limited devices

Unreviewed gardening.

  • stress/big-wasm-memory-grow-no-max.js:
  • stress/big-wasm-memory-grow.js:
  • stress/big-wasm-memory.js:
5:16 AM Changeset in webkit [238372] by commit-queue@webkit.org
  • 10 edits in trunk

[GTK][WPE] Solve some issues in the enumerateDevices code
​https://bugs.webkit.org/show_bug.cgi?id=191832

Patch by Alejandro G. Castro <​alex@igalia.com> on 2018-11-19
Reviewed by Carlos Garcia Campos.

Source/WebKit:

  • UIProcess/API/glib/WebKitDeviceInfoPermissionRequest.cpp:

Increase the version to 2.24.
(webkitDeviceInfoPermissionRequestAllow):
(webkitDeviceInfoPermissionRequestDeny):

  • UIProcess/API/glib/WebKitDeviceInfoPermissionRequestPrivate.h:

Replace the forward declaration with the include.

  • UIProcess/API/gtk/WebKitDeviceInfoPermissionRequest.h: Typo.
  • UIProcess/API/gtk/WebKitWebsiteData.h:

Increase the version to 2.24.

  • UIProcess/API/wpe/WebKitDeviceInfoPermissionRequest.h: Typo.
  • UIProcess/API/wpe/WebKitWebsiteData.h:

Increase the version to 2.24.

Tools:

  • TestWebKitAPI/Tests/WebKitGLib/TestWebsiteData.cpp:

(serverCallback): Make all the Strings with HTML code static.
(testWebsiteDataDeviceIdHashSalt): Change the webView settings in
the test.

  • TestWebKitAPI/glib/WebKitGLib/WebViewTest.cpp:

(WebViewTest::initializeWebView): Remove the initialization of the
webView settings, we now do it in the specific test.

1:16 AM Changeset in webkit [238371] by Carlos Garcia Campos
  • 4 edits in trunk/Source/WebKit

Unreviewed. Fix GTK+ build after r238318.

Add new symbols to API documentation files. They should have been added in r237031, but r238318 revealed it.

  • UIProcess/API/gtk/docs/webkit2gtk-4.0-sections.txt:
  • UIProcess/API/gtk/docs/webkit2gtk-4.0.types:
  • UIProcess/API/gtk/docs/webkit2gtk-docs.sgml:
1:13 AM Changeset in webkit [238370] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebKit

Regression(r238330): A lot of WebInspector tests are crashing on the bots
​https://bugs.webkit.org/show_bug.cgi?id=191814

Patch by Joseph Pecoraro <Joseph Pecoraro> on 2018-11-19
Reviewed by Ryosuke Niwa.

Expose the expected connection type from the UIProcess into the WebProcess.
When there is a local inspector connection to the UIProcess expose it as
a local frontend channel to the inspector controllers in the Page.

  • UIProcess/WebPageInspectorTargetAgent.h:
  • UIProcess/WebPageInspectorTargetAgent.cpp:

(WebKit::WebPageInspectorTargetAgent::frontendChannel):
An additional FrontendChannel class is not needed, just use self.

(WebKit::WebPageInspectorTargetAgent::WebPageInspectorTargetAgent):
(WebKit::WebPageInspectorTargetAgent::connectionType const):
(WebKit::WebPageInspectorTargetAgent::sendMessageToFrontend):
Expose the channel to targets as local if there is a local frontend
in the UIProcess.

Nov 18, 2018:

11:29 PM Changeset in webkit [238369] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit

Regression(r238330) ProcessSwap.WebInspector API test is a flaky crash in debug
​https://bugs.webkit.org/show_bug.cgi?id=191812

Patch by Joseph Pecoraro <Joseph Pecoraro> on 2018-11-18
Reviewed by Chris Dumez.

  • WebProcess/WebPage/WebInspector.cpp:

(WebKit::WebInspector::setFrontendConnection):
It can be possible to receive a frontend connection multiple times
if another WebProcess opened the inspector and this WebProcess got
swapped in before the InspectorProcess notified the WebProcess of
a connection to use. It is always best to use the latest connection
we have received, since any others would have been invalidated.

11:14 PM Changeset in webkit [238368] by Chris Dumez
  • 5 edits
    1 add in trunk

[PSON] Received an invalid message "WebPageProxy.DidPerformClientRedirect" from the web process
​https://bugs.webkit.org/show_bug.cgi?id=191828

Reviewed by Antti Koivisto.

Source/WebKit:

WebPageProxy::didPerformClientRedirect() had a MESSAGE_CHECK_URL() for both the source
and the destination URL. The macro calls WebProcessProxy::checkURLReceivedFromWebProcess()
to make sure that both URLs were actually loaded by this WebContent process.

However, now that we're enabled process-swap on cross-site navigation, these checks are no
longer fully valid. If the client-side redirect is cross-site, then the source and destination
URLs get loaded by 2 different WebContent processes. Since didPerformClientRedirect() gets
called after the redirect and gets set by the new WebContent process, only the destination URL
is guaranteed to have been loaded by this process. The source URL may have been loaded in
another process so this patch drops the message check for the source URL.

  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::didPerformClientRedirect):

Tools:

Add API test coverage.

  • TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
  • TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:
  • TestWebKitAPI/Tests/WebKitCocoa/client-side-redirect.html: Added.
10:51 PM Changeset in webkit [238367] by yusukesuzuki@slowstart.org
  • 7 edits in trunk/Source/JavaScriptCore

Consider removing double load for accessing the instructions from LLInt
​https://bugs.webkit.org/show_bug.cgi?id=190932

Reviewed by Mark Lam.

Changing InstructionStream to RefCountedArray like structure involves so much changes
including BytecodeGraph, PreciseJumpTargets etc. Instead, CodeBlock simply hold a raw
pointer to the InstructionStream's data. Since InstructionStream is not changed
anymore, this pointer is valid while CodeBlock is live.

  • bytecode/CodeBlock.cpp:

(JSC::CodeBlock::CodeBlock):

  • bytecode/CodeBlock.h:
  • bytecode/InstructionStream.h:

(JSC::InstructionStream::rawPointer const):

  • llint/LowLevelInterpreter.asm:
  • llint/LowLevelInterpreter32_64.asm:
  • llint/LowLevelInterpreter64.asm:
9:23 PM Changeset in webkit [238366] by Fujii Hironori
  • 3 edits in trunk/Source/JavaScriptCore

REGRESSION(r238039) WebCore::JSDOMGlobalObject::createStructure is using JSC::Structure::create without including StructureInlines.h
​https://bugs.webkit.org/show_bug.cgi?id=191626

Reviewed by Yusuke Suzuki.

JSC::Structure::create is used everywhere. It should be defined in
Structure.h, not in StructureInlines.h.

  • runtime/Structure.h:

(JSC::Structure::create): Moved.

  • runtime/StructureInlines.h: Moved JSC::Structure::create.
11:10 AM Changeset in webkit [238365] by yusukesuzuki@slowstart.org
  • 19 edits
    2 adds in trunk

Unreviewed, rolling in the rest of r237254
​https://bugs.webkit.org/show_bug.cgi?id=190340

JSTests:

  • ChakraCore/test/Function/FuncBodyES5.baseline-jsc:
  • stress/function-cache-with-parameters-end-position.js: Added.

(shouldBe):
(shouldThrow):
(i.anonymous):

  • stress/function-constructor-name.js: Added.

(shouldBe):
(GeneratorFunction):
(AsyncFunction.async):
(AsyncGeneratorFunction.async):
(anonymous):
(async.anonymous):

  • test262/expectations.yaml:

LayoutTests/imported/w3c:

  • web-platform-tests/html/webappapis/scripting/events/inline-event-handler-ordering-expected.txt:
  • web-platform-tests/html/webappapis/scripting/events/invalid-uncompiled-raw-handler-compiled-late-expected.txt:
  • web-platform-tests/html/webappapis/scripting/processing-model-2/compile-error-in-attribute-expected.txt:
  • web-platform-tests/html/webappapis/scripting/processing-model-2/compile-error-in-body-onerror-expected.txt:

Source/JavaScriptCore:

  • parser/ParserModes.h:
  • parser/ParserTokens.h:

(JSC::JSTextPosition::JSTextPosition):
(JSC::JSTokenLocation::JSTokenLocation): Deleted.

  • runtime/CodeCache.cpp:

(JSC::CodeCache::getUnlinkedGlobalFunctionExecutable):

  • runtime/FunctionConstructor.cpp:

(JSC::constructFunctionSkippingEvalEnabledCheck):

LayoutTests:

  • fast/dom/attribute-event-listener-errors-expected.txt:
  • fast/events/attribute-listener-deletion-crash-expected.txt:
  • fast/events/window-onerror-syntax-error-in-attr-expected.txt:
  • js/dom/invalid-syntax-for-function-expected.txt:
  • js/dom/script-start-end-locations-expected.txt:
10:20 AM Changeset in webkit [238364] by Michael Catanzaro
  • 3 edits in releases/WebKitGTK/webkit-2.22/Source/WebCore

Merge r238363 - HarfBuzzFace::CacheEntry should use 32-bit values in its HashMap
​https://bugs.webkit.org/show_bug.cgi?id=191825

Patch by Zan Dobersek <​zdobersek@igalia.com> on 2018-11-18
Reviewed by Michael Catanzaro.

The HashMap-based glyph cache contained in HarfBuzzFace::CacheEntry
objects is used to map given Unicode codepoints to corresponding
glyph indices (which occurs in the harfBuzzGetGlyph() function that's
invoked by HarfBuzz).

The 16-bit unsigned integer as the map's value type is not enough here
when the glyph index mapping is done through Freetype -- its API returns
32-bit glyph indices, and Cairo assigns this value to the 64-bit
unsigned index variable in the cairo_glyph_t struct. The value type is
thus bumped to 32 bits to match the unsigned type size of the index's
origin.

  • platform/graphics/harfbuzz/HarfBuzzFace.h:
  • platform/graphics/harfbuzz/HarfBuzzFaceCairo.cpp:
9:26 AM Changeset in webkit [238363] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebCore

HarfBuzzFace::CacheEntry should use 32-bit values in its HashMap
​https://bugs.webkit.org/show_bug.cgi?id=191825

Patch by Zan Dobersek <​zdobersek@igalia.com> on 2018-11-18
Reviewed by Michael Catanzaro.

The HashMap-based glyph cache contained in HarfBuzzFace::CacheEntry
objects is used to map given Unicode codepoints to corresponding
glyph indices (which occurs in the harfBuzzGetGlyph() function that's
invoked by HarfBuzz).

The 16-bit unsigned integer as the map's value type is not enough here
when the glyph index mapping is done through Freetype -- its API returns
32-bit glyph indices, and Cairo assigns this value to the 64-bit
unsigned index variable in the cairo_glyph_t struct. The value type is
thus bumped to 32 bits to match the unsigned type size of the index's
origin.

  • platform/graphics/harfbuzz/HarfBuzzFace.h:
  • platform/graphics/harfbuzz/HarfBuzzFaceCairo.cpp:
9:07 AM Changeset in webkit [238362] by Michael Catanzaro
  • 8 edits
    2 deletes in releases/WebKitGTK/webkit-2.22

Revert "Merge r237837 - [MediaStream] An audio track should be muted when capture is interrupted by the OS."

This reverts commit b8cbeaf8a2a641e7b2db25972dc6956778da8c2a.

This commit only affects Apple ports, so I shouldn't have backported it. It's harmless here and does not really need to be reverted, except this will serve as a useful reminder to myself that it was not needed here.

4:50 AM Changeset in webkit [238361] by yusukesuzuki@slowstart.org
  • 2 edits in trunk/Source/WebKit

Use Box<BinarySemaphore> instead of dispatch_semaphore_t
​https://bugs.webkit.org/show_bug.cgi?id=189691

Reviewed by Ryosuke Niwa.

Use Box<BinarySemaphore> instead, which utilizes our WTF::Lock mechanism.

  • NetworkProcess/watchos/NetworkProximityManager.mm:

(WebKit::NetworkProximityManager::updateRecommendation):

Note: See TracTimeline for information about the timeline view.