⚠ Archived content — this site is no longer maintained.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

Timeline



Jan 15, 2018:

11:03 PM Changeset in webkit [226963] by Matt Baker
  • 7 edits in trunk/Source/WebInspectorUI

Web Inspector: TabBar redesign: add context menu to TabBar for toggling available tabs
​https://bugs.webkit.org/show_bug.cgi?id=181448
<rdar://problem/36383298>

Reviewed by Devin Rousso.

  • UserInterface/Base/Main.js:

(WI.loaded):
(WI.contentLoaded):
Reorder production tab classes and default (open) tabs.
(WI.registerTabClass): Removed.
Not used.

  • UserInterface/Base/Object.js:

Remove notification that is no longer used.

  • UserInterface/Views/GeneralTabBarItem.js:

(WI.GeneralTabBarItem):
(WI.GeneralTabBarItem.prototype.set title):
(WI.GeneralTabBarItem.prototype._handleContextMenuEvent): Deleted.
Remove per-tab context menu (provided Close Tab and Close Other Tabs).

  • UserInterface/Views/NewTabContentView.js:

(WI.NewTabContentView):

  • UserInterface/Views/TabBar.js:

(WI.TabBar):
(WI.TabBar.prototype._handleContextMenu):

  • UserInterface/Views/TabBrowser.js:

(WI.TabBrowser):
(WI.TabBrowser._handleNewTabContextMenu): Deleted.
No longer needed.

10:44 PM Changeset in webkit [226962] by yoav@yoav.ws
  • 4 edits
    2 adds in trunk

Support for preconnect Link headers
​https://bugs.webkit.org/show_bug.cgi?id=181657

Reviewed by Darin Adler.

Source/WebCore:

Move the preconnect functionality into its own function, and
also call this function when Link headers are processed.

Test: http/tests/preconnect/link-header-rel-preconnect-http.php

  • loader/LinkLoader.cpp:

(WebCore::LinkLoader::loadLinksFromHeader): Call preconnect.
(WebCore::LinkLoader::preconnect): Preconnect to a host functionality moved here.
(WebCore::LinkLoader::preload): Renamed preloadIfNeeded to preload.
(WebCore::LinkLoader::loadLink): Call preconnect.

  • loader/LinkLoader.h:

LayoutTests:

Add test to see Link preconnect headers trigger a connection.

  • http/tests/preconnect/link-header-rel-preconnect-http-expected.txt: Added.
  • http/tests/preconnect/link-header-rel-preconnect-http.php: Added.
6:47 PM WebKitGTK/Gardening/Calendar edited by Michael Catanzaro
(diff)
6:46 PM WebKitGTK/Gardening/Calendar edited by Michael Catanzaro
(diff)
6:44 PM Changeset in webkit [226961] by Michael Catanzaro
  • 3 edits in trunk/LayoutTests

Unreviewed GTK layout test gardening

6:36 PM Changeset in webkit [226960] by rniwa@webkit.org
  • 2 edits in trunk/Websites/browserbench.org

ΩUpdate the default hyperlink for Speedometer to Speedometer 2.0.

Rubber-stamped by Saam Barati.

  • index.html:
6:20 PM Changeset in webkit [226959] by Michael Catanzaro
  • 2 edits in trunk/Tools

Unreviewed, add Thibault to contributors.json

It's a bit weird to add someone here before their first contribution, but this affects
Bugzilla autocomplete, and that is serious business.

  • Scripts/webkitpy/common/config/contributors.json:
3:49 PM Changeset in webkit [226958] by Michael Catanzaro
  • 14 edits in trunk/Source

Improve use of ExportMacros
​https://bugs.webkit.org/show_bug.cgi?id=181652

Reviewed by Konstantin Tokarev.

Source/bmalloc:

Disable BEXPORT on Linux ports.

  • bmalloc/BExport.h: Check for BUSE(EXPORT_MACROS).
  • bmalloc/BPlatform.h: Add BUSE(EXPORT_MACROS) and define it on macOS and iOS.

Source/JavaScriptCore:

  • API/JSBase.h: Update a comment.
  • inspector/InspectorBackendDispatcher.h: Use a better, yet equivalent, WTF macro.
  • runtime/JSExportMacros.h: Simplify the #defines in this file.

Source/WebCore:

Remove a comment.

  • platform/PlatformExportMacros.h:

Source/WebCore/PAL:

Remove a comment.

  • pal/ExportMacros.h:

Source/WTF:

  • wtf/ExportMacros.h: Simplify the #defines in this file.
  • wtf/Platform.h: Remove unneeded define. Remove comment that expects all ports to eventually enable the export macros. WPE will never want these. We don't currently want them for GTK either, though how we link GTK has been in flux recently.
2:56 PM Changeset in webkit [226957] by Konstantin Tokarev
  • 3 edits
    3 adds in trunk

image-rendering should affect scaling of border-image
​https://bugs.webkit.org/show_bug.cgi?id=169440

Reviewed by Michael Catanzaro.

Source/WebCore:

Test: fast/borders/border-image-pixelated.html

  • rendering/style/NinePieceImage.cpp:

(WebCore::NinePieceImage::paint):

LayoutTests:

  • fast/borders/border-image-pixelated-expected.html: Added.
  • fast/borders/border-image-pixelated.html: Added.
  • fast/borders/resources/border-image-pixelated.png: Added.
2:02 PM Changeset in webkit [226956] by jfbastien@apple.com
  • 9 edits in trunk

Remove makePoisonedUnique
​https://bugs.webkit.org/show_bug.cgi?id=181630
<rdar://problem/36498623>

Reviewed by Mark Lam.

Source/JavaScriptCore:

I added a conversion from std::unique_ptr, so we can just use
std::make_unique and it'll auto-poison when converted.

  • bytecode/CodeBlock.h:

(JSC::CodeBlock::makePoisonedUnique): Deleted.

  • runtime/JSGlobalObject.cpp:

(JSC::JSGlobalObject::init):

  • runtime/JSGlobalObject.h:

(JSC::JSGlobalObject::makePoisonedUnique): Deleted.

Source/WTF:

I added a conversion from std::unique_ptr, so we can just use
std::make_unique and it'll auto-poison when converted.

  • wtf/PoisonedUniquePtr.h:

(WTF::makePoisonedUnique): Deleted.

Tools:

Update tests.

  • TestWebKitAPI/Tests/WTF/PoisonedUniquePtr.cpp:

(TestWebKitAPI::TEST):

  • TestWebKitAPI/Tests/WTF/PoisonedUniquePtrForNonTriviallyDestructibleArrays.cpp:

(TestWebKitAPI::TEST):

1:24 PM Changeset in webkit [226955] by tpopela@redhat.com
  • 2 edits in trunk/Source/WebCore

2.19.3 ACCELERATED_2D_CANVAS support is broken
​https://bugs.webkit.org/show_bug.cgi?id=180799

Reviewed by Michael Catanzaro.

  • html/canvas/CanvasRenderingContext2DBase.cpp:

(WebCore::CanvasRenderingContext2DBase::didDraw):

1:21 PM Changeset in webkit [226954] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebCore

[Curl] Enable HTTP/2
​https://bugs.webkit.org/show_bug.cgi?id=181551

Patch by Basuke Suzuki <Basuke Suzuki> on 2018-01-15
Reviewed by Michael Catanzaro.

Start supporting HTTP/2 protocol. The first step is just enabling the HTTP/2 on Curl backend.
Next step will be to enable multiplexing feature.

No new tests because we don't have HTTP/2 test backend yet.

  • platform/network/curl/CurlContext.cpp:

(WebCore::CurlContext::isHttp2Enabled const):
(WebCore::CurlHandle::enableHttp):
(WebCore::CurlHandle::enableHttpGetRequest):
(WebCore::CurlHandle::enableHttpHeadRequest):
(WebCore::CurlHandle::enableHttpPostRequest):
(WebCore::CurlHandle::enableHttpPutRequest):
(WebCore::CurlHandle::setHttpCustomRequest):

  • platform/network/curl/CurlContext.h:
12:43 PM Changeset in webkit [226953] by dino@apple.com
  • 5 edits in trunk/Source/WebCore

Use a helper function for checked arithmetic in WebGL validation
​https://bugs.webkit.org/show_bug.cgi?id=181620
<rdar://problem/36485879>

Reviewed by Eric Carlson.

Eric recommended using a templated helper function to do
a common arithmetic check in WebGL validation.

  • html/canvas/WebGL2RenderingContext.cpp:

(WebCore::WebGL2RenderingContext::validateIndexArrayConservative):

  • html/canvas/WebGLRenderingContext.cpp:

(WebCore::WebGLRenderingContext::validateIndexArrayConservative):

  • html/canvas/WebGLRenderingContextBase.cpp:

(WebCore::WebGLRenderingContextBase::validateIndexArrayPrecise):
(WebCore::WebGLRenderingContextBase::validateDrawArrays):
(WebCore::WebGLRenderingContextBase::validateSimulatedVertexAttrib0):
(WebCore::WebGLRenderingContextBase::simulateVertexAttrib0):

  • html/canvas/WebGLRenderingContextBase.h:

(WebCore::WebGLRenderingContextBase::checkedAddAndMultiply): New helper.

11:16 AM Changeset in webkit [226952] by dino@apple.com
  • 6 edits in trunk/Source/WebCore

Use traits for animation timing functions
​https://bugs.webkit.org/show_bug.cgi?id=181651
<rdar://problem/36525328>

Reviewed by Antoine Quint.

Use the type traits for TimingFunction classes, so
we can is<> and downcast<>.

  • css/CSSComputedStyleDeclaration.cpp:

(WebCore::createTimingFunctionValue):

  • platform/animation/TimingFunction.cpp:

(WebCore::TimingFunction::transformTime const):

  • platform/animation/TimingFunction.h:
  • platform/graphics/ca/GraphicsLayerCA.cpp:

(WebCore::animationHasStepsTimingFunction):
(WebCore::animationHasFramesTimingFunction):

  • platform/graphics/ca/cocoa/PlatformCAAnimationCocoa.mm:

(WebCore::toCAMediaTimingFunction):

9:54 AM Changeset in webkit [226951] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebCore

RealtimeMediaSource should be ThreadSafeRefCounted
​https://bugs.webkit.org/show_bug.cgi?id=181649

Patch by Youenn Fablet <​youenn@apple.com> on 2018-01-15
Reviewed by Eric Carlson.

Difficult to write a test as this is really racy.
RealtimeIncomingVideoSourceCocoa::OnFrame is taking a reference on a background thread
to send a task to the main thread.
This requires it to be thread safe ref counted.

  • platform/mediastream/RealtimeMediaSource.h:
9:36 AM Changeset in webkit [226950] by Adrian Perez de Castro
  • 3 edits in trunk/Tools

[GTK][JHBuild] Tools/gtk/install-dependencies: Package libxfont does not exist anymore in Arch Linux
​https://bugs.webkit.org/show_bug.cgi?id=181476

Reviewed by Michael Catanzaro.

Update to xorg-server 1.19.6 which depends on libxfont2, and change
the lists of packages to install libxfont2-devel (or equivalent).

  • gtk/install-dependencies:
  • gtk/jhbuild.modules:
7:49 AM Changeset in webkit [226949] by Philippe Normand
  • 2 edits in trunk/LayoutTests

Unreviewed GTK gardening

  • platform/gtk/TestExpectations: Skip

media/media-fullscreen-pause-inline.html as advised by Eric
Carlson.

7:27 AM Changeset in webkit [226948] by Philippe Normand
  • 5 edits in trunk

Prevent useless MediaPlayer mute state notifications
​https://bugs.webkit.org/show_bug.cgi?id=181646

Reviewed by Carlos Garcia Campos.

Source/WebCore:

On GTK port the mute change notification was triggering the test
runner to think the whole page mute state had changed and that
media elements were muted. The simplest solution is to propagate
the notification only if the state actually changed.

  • platform/graphics/MediaPlayer.cpp:

(WebCore::MediaPlayer::muteChanged): Propagate mute notification
only if the mute state actually changed.

LayoutTests:

Unflag tests depending on muted state and media session
restrictions, most of those are now passing.

  • platform/gtk/TestExpectations:
  • platform/gtk/media/video-controls-no-scripting-expected.txt:
2:50 AM Changeset in webkit [226947] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebCore

[GStreamer] Don't wait for draw condition variable when shutting down.
​https://bugs.webkit.org/show_bug.cgi?id=180978

Patch by Sebastian Dröge <​sebastian@centricular.com> on 2018-01-15
Reviewed by Carlos Garcia Campos.

  • platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.cpp:

(WebCore::MediaPlayerPrivateGStreamerBase::triggerRepaint):
(WebCore::MediaPlayerPrivateGStreamerBase::cancelRepaint):

  • platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.h:

By also waiting for the draw condition variable while shutting down,
it is possible that the GStreamer video sink is waiting for the main
thread to actually render the current frame, while at the same time
the main thread is waiting for the GStreamer video sink to shut down,
resulting in a deadlock.

1:36 AM Changeset in webkit [226946] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit

ASSERTION FAILED: m_ptr under WebKit::CacheStorage::Caches::writeRecord
​https://bugs.webkit.org/show_bug.cgi?id=181401
<rdar://problem/36379022>

Patch by Youenn Fablet <​youenn@apple.com> on 2018-01-15
Reviewed by Darin Adler.

m_isInitialized is true but m_storage is set to nullptr when calling writeRecord.
The only case seems to be if:

  • We are doing persistent storage
  • We are traversing the storage to do the initialization. At that point m_storage is set to a correct value.
  • clearMemoryRepresentation is called, thus setting m_storage to nullptr and m_isInitialized to false.

We fix this by making clearMemoryRepresentation a no-op if we are initializing the caches.
clearMemoryRepresentation is about cleaning the in-memory information of the caches and the memory representation is empty at init time.
Nullifying m_storage is a memory consumption optimization.

  • NetworkProcess/cache/CacheStorageEngineCaches.cpp:

(WebKit::CacheStorage::Caches::clearMemoryRepresentation):

12:45 AM Changeset in webkit [226945] by Michael Catanzaro
  • 5 edits
    1 delete in trunk

REGRESSION(r226266): [GTK] RELEASE_ASSERT(reservedZoneSize >= minimumReservedZoneSize) in JSC::VM::updateStackLimits
​https://bugs.webkit.org/show_bug.cgi?id=181438
<rdar://problem/36376724>

Reviewed by Carlos Garcia Campos.

.:

Build JSC as a shared library.

Stop using -fvisibility=hidden. This is a shame, but it is causing problems that I don't
know how to fix.

  • Source/cmake/OptionsGTK.cmake:
  • Source/cmake/WebKitCompilerFlags.cmake:

Source/JavaScriptCore:

Roll out the functional changes of r226266. We'll keep the minor CMake library type setting
cleanup, but we have to switch back to building JSC only as a shared library, and we have to
get rid of the version script.

  • PlatformGTK.cmake:
  • javascriptcoregtk-symbols.map: Removed.
12:17 AM Changeset in webkit [226944] by Carlos Garcia Campos
  • 4 edits
    3 adds in trunk/Tools

[GTK][WPE] Add support for unit test expectations
​https://bugs.webkit.org/show_bug.cgi?id=181589

Reviewed by Michael Catanzaro.

We currently have a way to skip tests by annotating them in the api test runner script. The main problem of this
approach is that we skip tests when they fail in the bots and we never notice if they stop failing, keeping the
tests skipped forever. This is indeed the case of several WebKit2 C API tests. Annotating skipped tests in the
script itself is not a good idea either.

This patch adds a generic TestExpectations class for simple tests based on tests with subtests, like our unit
tests, but also WebDriver tests. It parses a json file with the tests and subtests expectations and provides
convenient methods to query them.

  • Scripts/run-gtk-tests:

(GtkTestRunner): Remove all Skipped and Slow tests marked here.

  • Scripts/run-wpe-tests:

(WPETestRunner): Ditto.

  • Scripts/webkitpy/common/test_expectations.py: Added.

(TestExpectations):
(TestExpectations.init):
(TestExpectations._port_name_for_expected):
(TestExpectations._expected_value):
(TestExpectations.skipped_tests):
(TestExpectations.skipped_subtests):
(TestExpectations._expectation_value):
(TestExpectations.is_slow):
(TestExpectations.get_expectation):

  • Scripts/webkitpy/common/test_expectations_unittest.py: Added.

(MockTestExpectations):
(MockTestExpectations.init):
(MockTestExpectations.is_skip):
(ExpectationsTest):
(assert_exp):
(assert_not_exp):
(assert_bad_exp):
(assert_skip):
(test_basic):
(test_skip):
(test_flaky):
(test_build_type):

  • TestWebKitAPI/glib/TestExpectations.json: Added.
  • glib/api_test_runner.py:

(TestRunner): Remove SkippedTest implementation.
(TestRunner.init): Create a TestExpectations.
(TestRunner._test_cases_to_skip): Use TestExpectations to check skipped tests.
(TestRunner._should_run_test_program): Ditto.
(TestRunner._run_test_glib): Use TestExpectations to check if test suite is slow.
(TestRunner._run_test_glib.parse_line.set_test_result): Register also tests passing.
(TestRunner._run_google_test): Use TestExpectations to check if test cases is slow and register tests passing.
(TestRunner.run_tests): Check if actual result is the expected one and register also unexpected passes.
(TestRunner.run_tests.report): Helper to write report to stdout.

Jan 14, 2018:

3:02 PM Changeset in webkit [226943] by mmaxfield@apple.com
  • 2 edits in trunk/LayoutTests

[macOS] [WK2] Unskip some consistently-passing tests
​https://bugs.webkit.org/show_bug.cgi?id=181226
<rdar://problem/23023848>

Reviewed by Simon Fraser.

According to the flakiness dashboard, these tests consistently pass now.

  • platform/mac-wk2/TestExpectations:
11:43 AM Changeset in webkit [226942] by sbarati@apple.com
  • 2 edits in trunk/Source/JavaScriptCore

Unreviewed. r226928 broke the CLOOP build. This patch fixes the CLOOP build.

  • bytecode/CallLinkStatus.cpp:

(JSC::CallLinkStatus::computeFromLLInt):
(JSC::CallLinkStatus::computeExitSiteData):

Note: See TracTimeline for information about the timeline view.