WebKit-https.git
4 months ago[Win] DumpRenderTreeWatchDog.py is not checked into source
pvollan@apple.com [Thu, 4 Apr 2019 17:39:48 +0000 (17:39 +0000)]
[Win] DumpRenderTreeWatchDog.py is not checked into source
https://bugs.webkit.org/show_bug.cgi?id=196567
<rdar://problem/49525901>

Reviewed by Lucas Forschler.

This script is used on Win EWS bots to terminate hanging DRT processes.

* EWSTools/dumprendertreewatchdog.py: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243879 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[macOS WK1] Layout Test webanimations/accelerated-transition-interrupted-on-composite...
commit-queue@webkit.org [Thu, 4 Apr 2019 17:39:32 +0000 (17:39 +0000)]
[macOS WK1] Layout Test webanimations/accelerated-transition-interrupted-on-composited-element.html is a flaky image failure
https://bugs.webkit.org/show_bug.cgi?id=189678
<rdar://problem/44621674>

Patch by Antoine Quint <graouts@apple.com> on 2019-04-04
Reviewed by Dean Jackson.

We make the test sturdier by waiting for two rAFs instead of just one to wait until animations have been updated (one rAF)
and accelerated animations have been committed (two rAFs).

We also use the "transitioncancel" event to determine the transition has been interrupted rather than using a rAF callback.

* webanimations/accelerated-transition-interrupted-on-composited-element.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243878 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoTemporarily disable delayed WebProcess launch on iOS to address PLT regression
cdumez@apple.com [Thu, 4 Apr 2019 17:36:20 +0000 (17:36 +0000)]
Temporarily disable delayed WebProcess launch on iOS to address PLT regression
https://bugs.webkit.org/show_bug.cgi?id=196604
<rdar://problem/49074131>

Reviewed by Geoffrey Garen.

Temporarily disable delayed WebProcess launch on iOS to address PLT regression. This
is while I figure out how to enable this on iOS without regressed PLT.

* UIProcess/WebProcessPool.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243877 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[JSC] We should consider moving UnlinkedFunctionExecutable::m_parentScopeTDZVariables...
ticaiolima@gmail.com [Thu, 4 Apr 2019 17:21:15 +0000 (17:21 +0000)]
[JSC] We should consider moving UnlinkedFunctionExecutable::m_parentScopeTDZVariables to RareData
https://bugs.webkit.org/show_bug.cgi?id=194944

Reviewed by Keith Miller.

JSTests:

* stress/verify-bytecode-generator-cached-variables-under-tdz.js: Added.

Source/JavaScriptCore:

Based on profile data collected on JetStream2, Speedometer 2 and
other benchmarks, it is very rare having non-empty
UnlinkedFunctionExecutable::m_parentScopeTDZVariables.

- Data collected from Speedometer2
    Total number of UnlinkedFunctionExecutable: 39463
    Total number of non-empty parentScopeTDZVars: 428 (~1%)

- Data collected from JetStream2
    Total number of UnlinkedFunctionExecutable: 83715
    Total number of non-empty parentScopeTDZVars: 5285 (~6%)

We also collected numbers on 6 of top 10 Alexia sites.

- Data collected from youtube.com
    Total number of UnlinkedFunctionExecutable: 29599
    Total number of non-empty parentScopeTDZVars: 97 (~0.3%)

- Data collected from twitter.com
    Total number of UnlinkedFunctionExecutable: 23774
    Total number of non-empty parentScopeTDZVars: 172 (~0.7%)

- Data collected from google.com
    Total number of UnlinkedFunctionExecutable: 33209
    Total number of non-empty parentScopeTDZVars: 174 (~0.5%)

- Data collected from amazon.com:
    Total number of UnlinkedFunctionExecutable: 15182
    Total number of non-empty parentScopeTDZVars: 166 (~1%)

- Data collected from facebook.com:
    Total number of UnlinkedFunctionExecutable: 54443
    Total number of non-empty parentScopeTDZVars: 269 (~0.4%)

- Data collected from netflix.com:
    Total number of UnlinkedFunctionExecutable: 39266
    Total number of non-empty parentScopeTDZVars: 97 (~0.2%)

Considering such numbers, this patch is moving `m_parentScopeTDZVariables`
to RareData. This decreases sizeof(UnlinkedFunctionExecutable) by
16 bytes. With this change, now UnlinkedFunctionExecutable constructors
receives an `Optional<VariableEnvironmentMap::Handle>` and only stores
it when `value != WTF::nullopt`. We also changed
UnlinkedFunctionExecutable::parentScopeTDZVariables() and it returns
`VariableEnvironment()` whenever the Executable doesn't have RareData,
or VariableEnvironmentMap::Handle is unitialized. This is required
because RareData is instantiated when any of its field is stored and
we can have an unitialized `Handle` even on cases when parentScopeTDZVariables
is `WTF::nullopt`.

Results on memory usage on JetStrem2 is neutral.

    Mean of memory peak on ToT: 4258633728 bytes (confidence interval: 249720072.95)
    Mean of memory peak on Changes: 4367325184 bytes (confidence interval: 321285583.61)

* builtins/BuiltinExecutables.cpp:
(JSC::BuiltinExecutables::createExecutable):
* bytecode/UnlinkedFunctionExecutable.cpp:
(JSC::UnlinkedFunctionExecutable::UnlinkedFunctionExecutable):
* bytecode/UnlinkedFunctionExecutable.h:
* bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::getVariablesUnderTDZ):

BytecodeGenerator::getVariablesUnderTDZ now also caches if m_cachedVariablesUnderTDZ
is empty, so we can properly return `WTF::nullopt` without the
reconstruction of a VariableEnvironment to check if it is empty.

* bytecompiler/BytecodeGenerator.h:
(JSC::BytecodeGenerator::makeFunction):
* parser/VariableEnvironment.h:
(JSC::VariableEnvironment::isEmpty const):
* runtime/CachedTypes.cpp:
(JSC::CachedCompactVariableMapHandle::decode const):

It returns an unitialized Handle when there is no
CompactVariableEnvironment. This can happen when RareData is ensured
because of another field.

(JSC::CachedFunctionExecutableRareData::encode):
(JSC::CachedFunctionExecutableRareData::decode const):
(JSC::CachedFunctionExecutable::encode):
(JSC::CachedFunctionExecutable::decode const):
(JSC::UnlinkedFunctionExecutable::UnlinkedFunctionExecutable):
* runtime/CodeCache.cpp:

Instead of creating a dummyVariablesUnderTDZ, we simply pass
WTF::nullopt.

(JSC::CodeCache::getUnlinkedGlobalFunctionExecutable):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243875 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoLog the error if VideoProcessing library cannot be dlopen
youenn@apple.com [Thu, 4 Apr 2019 16:52:02 +0000 (16:52 +0000)]
Log the error if VideoProcessing library cannot be dlopen
https://bugs.webkit.org/show_bug.cgi?id=196609

Reviewed by Eric Carlson.

* Source/webrtc/sdk/WebKit/VideoProcessingSoftLink.cpp:
(webrtc::initVideoProcessingVPModuleInitialize):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243873 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[WPE] Unreviewed gardening
dpino@igalia.com [Thu, 4 Apr 2019 16:51:31 +0000 (16:51 +0000)]
[WPE] Unreviewed gardening

Remove tests that are expected to fail but pass.

* platform/wpe/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243872 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoREGRESSION (r236501?): Layout Test compositing/geometry/limit-layer-bounds-opacity...
commit-queue@webkit.org [Thu, 4 Apr 2019 16:38:50 +0000 (16:38 +0000)]
REGRESSION (r236501?): Layout Test compositing/geometry/limit-layer-bounds-opacity-transition.html is flaky
https://bugs.webkit.org/show_bug.cgi?id=190210
<rdar://problem/45045989>

Unreviewed.

The flakiness dashboard shows that this test hasn't been flaky for a few weeks, reverting the flaky expectation.

Patch by Antoine Quint <graouts@apple.com> on 2019-04-04

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243871 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoUnreviewed, roll out r243858 which made tests time out.
cdumez@apple.com [Thu, 4 Apr 2019 15:44:37 +0000 (15:44 +0000)]
Unreviewed, roll out r243858 which made tests time out.

Disable new API tests on Windows instead.

* TestWebKitAPI/Tests/WebKit/ReloadPageAfterCrash.cpp:
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243870 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoCache bytecode for jsc.cpp helpers and fix CachedStringImpl
tzagallo@apple.com [Thu, 4 Apr 2019 15:41:07 +0000 (15:41 +0000)]
Cache bytecode for jsc.cpp helpers and fix CachedStringImpl
https://bugs.webkit.org/show_bug.cgi?id=196409

Reviewed by Saam Barati.

JSTests:

* stress/bytecode-cache-cached-string-impl.js: Added.
(f):
(g):
* stress/bytecode-cache-run-string.js: Added.

Source/JavaScriptCore:

Some of the helpers in jsc.cpp, such as `functionRunString`, were stll using
using `makeSource` instead of `jscSource`, which does not use the ShellSourceProvider
and therefore does not write the bytecode cache to disk.

Changing that revealed a bug in bytecode cache. The Encoder keeps a mapping
of pointers to offsets of already cached objects, in order to avoid caching
the same object twice. Similarly, the Decoder keeps a mapping from offsets
to pointers, in order to avoid creating multiple objects in memory for the
same cached object. The following was happening:
1) A StringImpl* S was cached as CachedPtr<CachedStringImpl> at offset O. We add
an entry in the Encoder mapping that S has already been encoded at O.
2) We cache StringImpl* S again, but now as CachedPtr<CachedUniquedStringImpl>.
We find an entry in the Encoder mapping for S, and return the offset O. However,
the object cached at O is a CachedPtr<CachedStringImpl> (i.e. not Uniqued).

3) When decoding, there are 2 possibilities:
3.1) We find S for the first time through a CachedPtr<CachedStringImpl>. In
this case, everything works as expected since we add an entry in the decoder
mapping from the offset O to the decoded StringImpl* S. The next time we find
S through the uniqued version, we'll return the already decoded S.
3.2) We find S through a CachedPtr<CachedUniquedStringImpl>. Now we have a
problem, since the CachedPtr has the offset of a CachedStringImpl (not uniqued),
which has a different shape and we crash.

We fix this by making CachedStringImpl and CachedUniquedStringImpl share the
same implementation. Since it doesn't matter whether a string is uniqued for
encoding, and we always decode strings as uniqued either way, they can be used
interchangeably.

* jsc.cpp:
(functionRunString):
(functionLoadString):
(functionDollarAgentStart):
(functionCheckModuleSyntax):
(runInteractive):
* runtime/CachedTypes.cpp:
(JSC::CachedUniquedStringImplBase::decode const):
(JSC::CachedFunctionExecutable::rareData const):
(JSC::CachedCodeBlock::rareData const):
(JSC::CachedFunctionExecutable::encode):
(JSC::CachedCodeBlock<CodeBlockType>::encode):
(JSC::CachedUniquedStringImpl::encode): Deleted.
(JSC::CachedUniquedStringImpl::decode const): Deleted.
(JSC::CachedStringImpl::encode): Deleted.
(JSC::CachedStringImpl::decode const): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243869 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[Web Animations] JS wrapper may be deleted while animation is yet to dispatch its...
commit-queue@webkit.org [Thu, 4 Apr 2019 14:16:05 +0000 (14:16 +0000)]
[Web Animations] JS wrapper may be deleted while animation is yet to dispatch its finish event
https://bugs.webkit.org/show_bug.cgi?id=196118
<rdar://problem/46614137>

Patch by Antoine Quint <graouts@apple.com> on 2019-04-04
Reviewed by Ryosuke Niwa.

Source/WebCore:

Test: webanimations/js-wrapper-kept-alive.html

We need to teach WebAnimation to keep its JS wrapper alive if it's relevant or could become relevant again by virtue of having a timeline.

* animation/WebAnimation.cpp:
(WebCore::WebAnimation::stop): Drive-by fix for the missing superclass method call.
(WebCore::WebAnimation::hasPendingActivity const):
* animation/WebAnimation.h:

LayoutTests:

Add a test that starts a short animation, sets a custom property on it, registers a "finish" event listener on it and deletes
the sole reference to it in the JS world before triggering garbage collection. Prior to this fix, this test would time out
because the JS wrapper would be garbage-collected prior to the animation completing and thus the event listener would not
be called. To complete successfully, this test checks that it receives the event and its target is the same animation object
that was originally created by checking the custom property is still set.

* legacy-animation-engine/animations/resume-after-page-cache.html:
* webanimations/js-wrapper-kept-alive-expected.txt: Added.
* webanimations/js-wrapper-kept-alive.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243868 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoUnlinkedCodeBlock constructor from cache should initialize m_didOptimize
tzagallo@apple.com [Thu, 4 Apr 2019 14:15:20 +0000 (14:15 +0000)]
UnlinkedCodeBlock constructor from cache should initialize m_didOptimize
https://bugs.webkit.org/show_bug.cgi?id=196396

Reviewed by Saam Barati.

The UnlinkedCodeBlock constructor in CachedTypes was missing the initialization
for m_didOptimize, which leads to crashes in CodeBlock::thresholdForJIT.

* runtime/CachedTypes.cpp:
(JSC::UnlinkedCodeBlock::UnlinkedCodeBlock):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243867 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[GTK][WPE] Use a timer to request the creation of pending tiles
magomez@igalia.com [Thu, 4 Apr 2019 12:51:43 +0000 (12:51 +0000)]
[GTK][WPE] Use a timer to request the creation of pending tiles
https://bugs.webkit.org/show_bug.cgi?id=196594

Reviewed by Žan Doberšek.

Use a timer to request pending tile creation, as calls to notifyFlushRequired() are discarded
while inside a layer flush.

* platform/graphics/texmap/coordinated/CoordinatedGraphicsLayer.cpp:
(WebCore::CoordinatedGraphicsLayer::CoordinatedGraphicsLayer):
(WebCore::CoordinatedGraphicsLayer::flushCompositingStateForThisLayerOnly):
(WebCore::CoordinatedGraphicsLayer::updateContentBuffers):
(WebCore::CoordinatedGraphicsLayer::requestPendingTileCreationTimerFired):
* platform/graphics/texmap/coordinated/CoordinatedGraphicsLayer.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243866 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[CoordGraphics] Revalidate need for 'coordinated update completion' in ThreadedCompositor
magomez@igalia.com [Thu, 4 Apr 2019 08:53:36 +0000 (08:53 +0000)]
[CoordGraphics] Revalidate need for 'coordinated update completion' in ThreadedCompositor
https://bugs.webkit.org/show_bug.cgi?id=188839

Reviewed by Žan Doberšek.

Even if a platform layer has changed, don't wait for a main thread callback to finish the
CompositingRunLoop update. It can be finished as soon as we receive the frameComplete signal.

* Shared/CoordinatedGraphics/threadedcompositor/CompositingRunLoop.cpp:
(WebKit::CompositingRunLoop::scheduleUpdate):
(WebKit::CompositingRunLoop::stopUpdates):
(WebKit::CompositingRunLoop::updateCompleted):
(WebKit::CompositingRunLoop::updateTimerFired):
(WebKit::CompositingRunLoop::compositionCompleted): Deleted.
* Shared/CoordinatedGraphics/threadedcompositor/CompositingRunLoop.h:
* Shared/CoordinatedGraphics/threadedcompositor/ThreadedCompositor.cpp:
(WebKit::ThreadedCompositor::renderLayerTree):
(WebKit::ThreadedCompositor::sceneUpdateFinished):
(WebKit::ThreadedCompositor::displayRefreshMonitor):
(WebKit::ThreadedCompositor::handleDisplayRefreshMonitorUpdate): Deleted.
* Shared/CoordinatedGraphics/threadedcompositor/ThreadedCompositor.h:
* WebProcess/WebPage/CoordinatedGraphics/LayerTreeHost.cpp:
(WebKit::LayerTreeHost::handleDisplayRefreshMonitorUpdate):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243865 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[ATK] Wrong index passed to AtkObject::children-changed::add signal in AXObjectCache...
carlosgc@webkit.org [Thu, 4 Apr 2019 05:53:30 +0000 (05:53 +0000)]
[ATK] Wrong index passed to AtkObject::children-changed::add signal in AXObjectCache::attachWrapper()
https://bugs.webkit.org/show_bug.cgi?id=196538

Reviewed by Michael Catanzaro.

In most of the cases the parent is not found, probably because the child is not a direct descendant of the
parent returned by parentObjectUnignored(). We need to handle the case of find() returning notFound.

* accessibility/atk/AXObjectCacheAtk.cpp:
(WebCore::AXObjectCache::attachWrapper): Use -1 as the index when find() returns notFound.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243864 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[ATK] Cleanup WebPageAccessibilityObjectAtk
carlosgc@webkit.org [Thu, 4 Apr 2019 05:52:44 +0000 (05:52 +0000)]
[ATK] Cleanup WebPageAccessibilityObjectAtk
https://bugs.webkit.org/show_bug.cgi?id=196537

Reviewed by Michael Catanzaro.

Source/WebKit:

Several changes and cleanups:

 - Add WebKit prefix so that style checker doesn't complain about GObject conventions.
 - Rename the header to remove the Atk prefix to match the cpp file and class name.
 - Use pragma once.
 - Use nullptr instead of 0.
 - Use WEBKIT_DEFINE_TYPE instead of G_DEFINE_TYPE.
 - Return generic AtkObject* from constructor.

* SourcesGTK.txt:
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/atk/WebKitWebPageAccessibilityObject.cpp: Renamed from Source/WebKit/WebProcess/WebPage/atk/WebPageAccessibilityObjectAtk.cpp.
(accessibilityRootObjectWrapper):
(webkitWebPageAccessibilityObjectInitialize):
(webkitWebPageAccessibilityObjectGetIndexInParent):
(webkitWebPageAccessibilityObjectGetNChildren):
(webkitWebPageAccessibilityObjectRefChild):
(webkit_web_page_accessibility_object_class_init):
(webkitWebPageAccessibilityObjectNew):
(webkitWebPageAccessibilityObjectRefresh):
* WebProcess/WebPage/atk/WebKitWebPageAccessibilityObject.h: Added.
* WebProcess/WebPage/atk/WebPageAccessibilityObject.h: Removed.
* WebProcess/WebPage/gtk/WebPageGtk.cpp:
(WebKit::WebPage::platformInitialize):
(WebKit::WebPage::updateAccessibilityTree):

Tools:

Also consider files under atk directories as exceptions for GObject conventions.

* Scripts/webkitpy/style/checkers/cpp.py:
(check_identifier_name_in_declaration):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243863 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[ATK] Set ATK_STATE_TRANSIENT on the atk socket until it's embedded
carlosgc@webkit.org [Thu, 4 Apr 2019 05:51:37 +0000 (05:51 +0000)]
[ATK] Set ATK_STATE_TRANSIENT on the atk socket until it's embedded
https://bugs.webkit.org/show_bug.cgi?id=196535

Reviewed by Michael Catanzaro.

This way the properties are not cached. Right now, if number of children is queried before the socket is
embedded, 0 is returned and cached, so ATs don't see the child even after the socket has been embedded.

* UIProcess/API/gtk/WebKitWebViewBaseAccessible.cpp:
(webkitWebViewBaseAccessibleRefStateSet): Set ATK_STATE_TRANSIENT is the socket is not occupied.
* UIProcess/gtk/WebPageProxyGtk.cpp:
(WebKit::WebPageProxy::bindAccessibilityTree): Unset ATK_STATE_TRANSIENT.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243862 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[ATK] Embed the AtkSocket as soon as we receive the plug ID
carlosgc@webkit.org [Thu, 4 Apr 2019 05:50:22 +0000 (05:50 +0000)]
[ATK] Embed the AtkSocket as soon as we receive the plug ID
https://bugs.webkit.org/show_bug.cgi?id=196534

Reviewed by Michael Catanzaro.

We are currently storing the ID and waiting for the next time get_accessible is called to embed the socket. We
can simply embed the socket when the plug ID is received.

* UIProcess/API/gtk/WebKitWebViewBase.cpp:
(webkitWebViewBaseGetAccessible):
* UIProcess/WebPageProxy.h:
* UIProcess/gtk/WebPageProxyGtk.cpp:
(WebKit::WebPageProxy::bindAccessibilityTree):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243861 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[ATK] Don't touch accessibility tree in WebFrameLoaderClient::dispatchDidClearWindowO...
carlosgc@webkit.org [Thu, 4 Apr 2019 05:48:20 +0000 (05:48 +0000)]
[ATK] Don't touch accessibility tree in WebFrameLoaderClient::dispatchDidClearWindowObjectInWorld
https://bugs.webkit.org/show_bug.cgi?id=193914

Reviewed by Michael Catanzaro.

Move it to dispatchDidFinishDocumentLoad, since we know we have the document at that point and we can create the
root accessibility object wrapper.

* WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:
(WebKit::WebFrameLoaderClient::dispatchDidFinishDocumentLoad):
(WebKit::WebFrameLoaderClient::dispatchDidClearWindowObjectInWorld):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243860 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoUpdate AutoFill field icons to be SVG instead of PNG images.
timothy@apple.com [Thu, 4 Apr 2019 05:47:44 +0000 (05:47 +0000)]
Update AutoFill field icons to be SVG instead of PNG images.
https://bugs.webkit.org/show_bug.cgi?id=196557
rdar://problem/48292514

Reviewed by Tim Horton.

Source/WebCore:

* css/html.css:
(input::-webkit-credentials-auto-fill-button):
(input::-webkit-contacts-auto-fill-button):
(input::-webkit-credit-card-auto-fill-button):

LayoutTests:

* platform/ios/fast/forms/auto-fill-button/input-auto-fill-button-expected.txt:
* platform/mac/fast/forms/auto-fill-button/input-contacts-auto-fill-button-expected.txt:
* platform/win/fast/forms/auto-fill-button/input-auto-fill-button-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243859 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoThe page's focusedFrame / frameSetLargestFrame do not get cleared on process swap...
Hironori.Fujii@sony.com [Thu, 4 Apr 2019 05:44:01 +0000 (05:44 +0000)]
The page's focusedFrame / frameSetLargestFrame do not get cleared on process swap or crash
https://bugs.webkit.org/show_bug.cgi?id=196588
<rdar://problem/49365787>

Unreviewed build fix for Windows port.

error C3861: 'kill': identifier not found

* TestWebKitAPI/Tests/WebKit/ReloadPageAfterCrash.cpp:
(TestWebKitAPI::TEST): Replaced kill with WKPageTerminate.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243858 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoUnreviewed, rolling in r243843 with the build fix
ysuzuki@apple.com [Thu, 4 Apr 2019 05:27:25 +0000 (05:27 +0000)]
Unreviewed, rolling in r243843 with the build fix
https://bugs.webkit.org/show_bug.cgi?id=196586

* runtime/Options.cpp:
(JSC::recomputeDependentOptions):
* runtime/Options.h:
* runtime/RandomizingFuzzerAgent.cpp:
(JSC::RandomizingFuzzerAgent::getPrediction):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243857 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoSimplify some "programmaticScroll" code paths
simon.fraser@apple.com [Thu, 4 Apr 2019 03:57:22 +0000 (03:57 +0000)]
Simplify some "programmaticScroll" code paths
https://bugs.webkit.org/show_bug.cgi?id=196589

Reviewed by Zalan Bujtas.

Source/WebCore:

AsyncScrollingCoordinator::scheduleUpdateScrollPositionAfterAsyncScroll() just returned early if programmaticScroll
was true, so instead, just never call it. This means we can remove the "programmaticScroll" argument from
scheduleUpdateScrollPositionAfterAsyncScroll(). Also change some callers to use the ScrollType enum
instead of a bool.

Now, ThreadedScrollingTree::scrollingTreeNodeDidScroll() just returns early. Programmatic scrolls
update state on the main thread before updating the scrolling tree, so this makes sense.

* page/scrolling/AsyncScrollingCoordinator.cpp:
(WebCore::AsyncScrollingCoordinator::requestScrollPositionUpdate):
(WebCore::AsyncScrollingCoordinator::scheduleUpdateScrollPositionAfterAsyncScroll):
(WebCore::AsyncScrollingCoordinator::updateScrollPositionAfterAsyncScrollTimerFired):
(WebCore::AsyncScrollingCoordinator::updateScrollPositionAfterAsyncScroll):
(WebCore::AsyncScrollingCoordinator::reconcileScrollingState):
* page/scrolling/AsyncScrollingCoordinator.h:
(WebCore::AsyncScrollingCoordinator::ScheduledScrollUpdate::ScheduledScrollUpdate):
(WebCore::AsyncScrollingCoordinator::ScheduledScrollUpdate::matchesUpdateType const):
* page/scrolling/ScrollingCoordinator.cpp:
(WebCore::operator<<):
* page/scrolling/ScrollingCoordinator.h:
(WebCore::ScrollingCoordinator::reconcileScrollingState):
* page/scrolling/ThreadedScrollingTree.cpp:
(WebCore::ThreadedScrollingTree::scrollingTreeNodeDidScroll):

Source/WebKit:

* WebProcess/WebPage/RemoteLayerTree/RemoteScrollingCoordinator.mm: Remove the parameter.
(WebKit::RemoteScrollingCoordinator::scrollPositionChangedForNode): Use the enum type.
* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::updateVisibleContentRects):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243855 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[ews-build] Cancel build and similar operations should have authentication
aakash_jain@apple.com [Thu, 4 Apr 2019 03:57:11 +0000 (03:57 +0000)]
[ews-build] Cancel build and similar operations should have authentication
https://bugs.webkit.org/show_bug.cgi?id=196520

Reviewed by Dewei Zhu.

* BuildSlaveSupport/ews-build/master.cfg: Configure authentication using admin's username and password from
environment variables.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243854 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoUnreviewed, rolling out r243843.
ryanhaddad@apple.com [Thu, 4 Apr 2019 03:44:18 +0000 (03:44 +0000)]
Unreviewed, rolling out r243843.

Broke CLoop and Windows builds.

Reverted changeset:

"[JSC] Add dump feature for RandomizingFuzzerAgent"
https://bugs.webkit.org/show_bug.cgi?id=196586
https://trac.webkit.org/changeset/243843

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243853 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[ews-build] Do not trigger queues which are not in production yet
aakash_jain@apple.com [Thu, 4 Apr 2019 03:38:01 +0000 (03:38 +0000)]
[ews-build] Do not trigger queues which are not in production yet
https://bugs.webkit.org/show_bug.cgi?id=196581

Reviewed by Dewei Zhu.

* BuildSlaveSupport/ews-build/config.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243852 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoB3 should use associativity to optimize expression trees
rmorisset@apple.com [Thu, 4 Apr 2019 03:37:23 +0000 (03:37 +0000)]
B3 should use associativity to optimize expression trees
https://bugs.webkit.org/show_bug.cgi?id=194081

Reviewed by Filip Pizlo.

JSTests:

Added three microbenchmarks:
- add-tree should be the ideal case, but there is no speedup because we are currently unable to prove that the CheckAdd won't overflow
- bit-xor-tree most closely matches the situation where the optimization triggers on the JetStream2 subtests where it triggers:
  an unbalanced expression tree of size 8 that can be balanced, with no other optimizations being unlocked. 16% speedup
- bit-or-tree is an ideal case, where the reassociation also enables a ton of further simplifications. 42% speedup

* microbenchmarks/add-tree.js: Added.
* microbenchmarks/bit-or-tree.js: Added.
* microbenchmarks/bit-xor-tree.js: Added.

Source/JavaScriptCore:

This patch adds a new B3 pass, that tries to find and optimize expression trees made purely of any one associative and commutative operator (Add/Mul/BitOr/BitAnd/BitXor).
The pass only runs in O2, and runs once, after lowerMacros and just before a run of B3ReduceStrength (which helps clean up the dead code it tends to leave behind).
I had to separate killDeadCode out of B3ReduceStrength (as a new B3EliminateDeadCode pass) to run it before B3OptimizeAssociativeExpressionTrees, as otherwise it is stopped by high use counts
inherited from CSE.
This extra run of DCE is by itself a win, most notably on microbenchmarks/instanceof-always-hit-two (1.5x faster), and on microbenchmarks/licm-dragons(-out-of-bounds) (both get 1.16x speedup).
I suspect it is because it runs between CSE and tail-dedup, and as a result allows a lot more tail-dedup to occur.

The pass is currently extremely conservative, not trying anything if it would cause _any_ code duplication.
For this purpose, it starts by computing use counts for the potentially interesting nodes (those with the right opcodes), and segregate them into expression trees.
The root of an expression tree is a node that is either used in multiple places, or is used by a value with a different opcode.
The leaves of an expression tree are nodes that are either used in multiple places, or have a different opcode.
All constant leaves of a tree are combined, as well as all leaves that are identical. What remains is then laid out into a balanced binary tree, hopefully maximizing ILP.

This optimization was implemented as a stand-alone pass and not as part of B3ReduceStrength mostly because it needs use counts to avoid code duplication.
It also benefits from finding all tree roots first, and not trying to repeatedly optimize subtrees.

I added several tests to testB3 with varying patterns of trees. It is also tested in a less focused way by lots of older tests.

In the future this pass could be expanded to allow some bounded amount of code duplication, and merging more leaves (e.g. Mul(a, 3) and a in an Add tree, into Mul(a, 4))
The latter will need exposing the peephole optimizations out of B3ReduceStrength to avoid duplicating code.

* JavaScriptCore.xcodeproj/project.pbxproj:
* Sources.txt:
* b3/B3Common.cpp:
(JSC::B3::shouldDumpIR):
(JSC::B3::shouldDumpIRAtEachPhase):
* b3/B3Common.h:
* b3/B3EliminateDeadCode.cpp: Added.
(JSC::B3::EliminateDeadCode::run):
(JSC::B3::eliminateDeadCode):
* b3/B3EliminateDeadCode.h: Added.
(JSC::B3::EliminateDeadCode::EliminateDeadCode):
* b3/B3Generate.cpp:
(JSC::B3::generateToAir):
* b3/B3OptimizeAssociativeExpressionTrees.cpp: Added.
(JSC::B3::OptimizeAssociativeExpressionTrees::OptimizeAssociativeExpressionTrees):
(JSC::B3::OptimizeAssociativeExpressionTrees::neutralElement):
(JSC::B3::OptimizeAssociativeExpressionTrees::isAbsorbingElement):
(JSC::B3::OptimizeAssociativeExpressionTrees::combineConstants):
(JSC::B3::OptimizeAssociativeExpressionTrees::emitValue):
(JSC::B3::OptimizeAssociativeExpressionTrees::optimizeRootedTree):
(JSC::B3::OptimizeAssociativeExpressionTrees::run):
(JSC::B3::optimizeAssociativeExpressionTrees):
* b3/B3OptimizeAssociativeExpressionTrees.h: Added.
* b3/B3ReduceStrength.cpp:
* b3/B3Value.cpp:
(JSC::B3::Value::replaceWithIdentity):
* b3/testb3.cpp:
(JSC::B3::testBitXorTreeArgs):
(JSC::B3::testBitXorTreeArgsEven):
(JSC::B3::testBitXorTreeArgImm):
(JSC::B3::testAddTreeArg32):
(JSC::B3::testMulTreeArg32):
(JSC::B3::testBitAndTreeArg32):
(JSC::B3::testBitOrTreeArg32):
(JSC::B3::run):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243851 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoAdd logging and ASSERTs to investigate issue with VPModuleInitialize
youenn@apple.com [Thu, 4 Apr 2019 03:34:53 +0000 (03:34 +0000)]
Add logging and ASSERTs to investigate issue with VPModuleInitialize
https://bugs.webkit.org/show_bug.cgi?id=196573

Reviewed by Eric Carlson.

Source/ThirdParty/libwebrtc:

Expand macros directly to add some logging.
Removed the dispatch_once since VPModuleInitialize is already called in one.

* Source/webrtc/sdk/WebKit/VideoProcessingSoftLink.cpp:
(webrtc::initVideoProcessingVPModuleInitialize):

Source/WebCore:

Add some ASSERTs.
No change of behavior.

* platform/mediastream/libwebrtc/LibWebRTCProviderCocoa.cpp:
(WebCore::LibWebRTCProviderCocoa::createDecoderFactory):
(WebCore::LibWebRTCProviderCocoa::createEncoderFactory):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243850 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoUnreviewed, fix typo in the name of API test added in r243848.
cdumez@apple.com [Thu, 4 Apr 2019 03:23:11 +0000 (03:23 +0000)]
Unreviewed, fix typo in the name of API test added in r243848.

* TestWebKitAPI/Tests/WebKit/ReloadPageAfterCrash.cpp:
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243849 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoThe page's focusedFrame / frameSetLargestFrame do not get cleared on process swap...
cdumez@apple.com [Thu, 4 Apr 2019 03:09:48 +0000 (03:09 +0000)]
The page's focusedFrame / frameSetLargestFrame do not get cleared on process swap or crash
https://bugs.webkit.org/show_bug.cgi?id=196588
<rdar://problem/49365787>

Reviewed by Ryosuke Niwa.

Source/WebKit:

The page's focusedFrame / frameSetLargestFrame do not get cleared on process swap or crash.
This can lead to returning stale frames to the client if it asks for those.

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

Tools:

Add API test coverage.

* TestWebKitAPI/Tests/WebKit/ReloadPageAfterCrash.cpp:
(TestWebKitAPI::nullJavaScriptCallback):
(TestWebKitAPI::didCrashCheckFrames):
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243848 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoRemove the isProgrammaticScroll argument to requestScroll() because no-one uses it
simon.fraser@apple.com [Thu, 4 Apr 2019 02:53:02 +0000 (02:53 +0000)]
Remove the isProgrammaticScroll argument to requestScroll() because no-one uses it
https://bugs.webkit.org/show_bug.cgi?id=196587

Reviewed by Zalan Bujtas.

This argument was unused.

* UIProcess/API/gtk/PageClientImpl.cpp:
(WebKit::PageClientImpl::requestScroll):
* UIProcess/API/gtk/PageClientImpl.h:
* UIProcess/API/wpe/PageClientImpl.cpp:
(WebKit::PageClientImpl::requestScroll):
* UIProcess/API/wpe/PageClientImpl.h:
* UIProcess/PageClient.h:
* UIProcess/RemoteLayerTree/RemoteLayerTreeDrawingAreaProxy.mm:
(WebKit::RemoteLayerTreeDrawingAreaProxy::commitLayerTree):
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::requestScroll):
* UIProcess/WebPageProxy.h:
* UIProcess/ios/PageClientImplIOS.h:
* UIProcess/ios/PageClientImplIOS.mm:
(WebKit::PageClientImpl::requestScroll):
* UIProcess/mac/PageClientImplMac.h:
* UIProcess/mac/PageClientImplMac.mm:
(WebKit::PageClientImpl::requestScroll):
* UIProcess/win/PageClientImpl.cpp:
(WebKit::PageClientImpl::requestScroll):
* UIProcess/win/PageClientImpl.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243847 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoRemove some redundant memebers from ScrollingStateFrameScrollingNode
simon.fraser@apple.com [Thu, 4 Apr 2019 02:52:58 +0000 (02:52 +0000)]
Remove some redundant memebers from ScrollingStateFrameScrollingNode
https://bugs.webkit.org/show_bug.cgi?id=196571

Reviewed by Zalan Bujtas.

m_requestedScrollPosition and m_requestedScrollPositionRepresentsProgrammaticScroll were
duplicated on ScrollingStateFrameScrollingNode and ScrollingStateScrollingNode, so
remove them from the derived class.

* page/scrolling/ScrollingStateFrameScrollingNode.cpp:
(WebCore::ScrollingStateFrameScrollingNode::ScrollingStateFrameScrollingNode):
(WebCore::ScrollingStateFrameScrollingNode::dumpProperties const):
* page/scrolling/ScrollingStateFrameScrollingNode.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243846 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoREGRESSION (r220717): Assertion fires when animating an SVG rounded corner rect till...
commit-queue@webkit.org [Thu, 4 Apr 2019 02:09:50 +0000 (02:09 +0000)]
REGRESSION (r220717): Assertion fires when animating an SVG rounded corner rect till it collapses
https://bugs.webkit.org/show_bug.cgi?id=196518

Patch by Said Abou-Hallawa <sabouhallawa@apple.com> on 2019-04-03
Reviewed by Simon Fraser.

Source/WebCore:

r220717 made RenderSVGRect clear its m_path in updateShapeFromElement().

RenderSVGRect tries to optimize its layout and drawing if the rectangle
is not rounded. So it uses the flag m_usePathFallback to know whether to
use m_path or m_innerStrokeRect and m_outerStrokeRect. If the rectangle
is rounded but its boundingSize is empty, m_path will be cleared,
m_innerStrokeRect and m_outerStrokeRect will be recalculated but
m_usePathFallback will not be reset to false. Therefore when calling
RenderSVGRect::isEmpty(), it will call RenderSVGShape::isEmpty() which
will assert since m_path is null.

Test: svg/animations/animate-rounded-corner-rect-zero-height.svg

* rendering/svg/RenderSVGRect.cpp:
(WebCore::RenderSVGRect::updateShapeFromElement):
Reset m_usePathFallback to false once clearPath() is called.

LayoutTests:

* svg/animations/animate-rounded-corner-rect-zero-height-expected.txt: Added.
* svg/animations/animate-rounded-corner-rect-zero-height.svg: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243845 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoNullptr crash in InlineTextBox::selectionState via TextIndicator::createWithRange
rniwa@webkit.org [Thu, 4 Apr 2019 02:01:07 +0000 (02:01 +0000)]
Nullptr crash in InlineTextBox::selectionState via TextIndicator::createWithRange
https://bugs.webkit.org/show_bug.cgi?id=196579

Reviewed by Simon Fraser.

Avoid crashing accessing the unengated optional's value in relese builds for now.

Unfortunately, fixing the underlying cause of the selection states of RenderView & RenderObject
getting out out of sync would require a significant re-architecturing of the whole selection
repainting / state managing mechanism.

* rendering/SelectionRangeData.h:
(WebCore::SelectionRangeData::startPosition const):
(WebCore::SelectionRangeData::endPosition const):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243844 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[JSC] Add dump feature for RandomizingFuzzerAgent
ysuzuki@apple.com [Thu, 4 Apr 2019 01:28:49 +0000 (01:28 +0000)]
[JSC] Add dump feature for RandomizingFuzzerAgent
https://bugs.webkit.org/show_bug.cgi?id=196586

Reviewed by Saam Barati.

Towards deterministic tests for the results from randomizing fuzzer agent, this patch adds Options::dumpRandomizingFuzzerAgentPredictions, which dumps the generated types.
The results is like this.

    getPrediction name:(#C2q9xD),bytecodeIndex:(22),original:(Array),generated:(OtherObj|Array|Float64Array|BigInt|NonIntAsDouble)
    getPrediction name:(makeUnwriteableUnconfigurableObject#AiEJv1),bytecodeIndex:(14),original:(OtherObj),generated:(Final|Uint8Array|Float64Array|SetObject|WeakSetObject|BigInt|NonIntAsDouble)

* runtime/Options.cpp:
(JSC::recomputeDependentOptions):
* runtime/Options.h:
* runtime/RandomizingFuzzerAgent.cpp:
(JSC::RandomizingFuzzerAgent::getPrediction):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243843 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agorun-webkit-tests: Upload test results (new results database)
jbedard@apple.com [Thu, 4 Apr 2019 00:43:05 +0000 (00:43 +0000)]
run-webkit-tests: Upload test results (new results database)
https://bugs.webkit.org/show_bug.cgi?id=196577
<rdar://problem/34841155>

Reviewed by Lucas Forschler.

* Scripts/webkitpy/layout_tests/controllers/manager.py:
(Manager.run): Upload results after each device shard, since the configuration will be
different for each device.
(Manager._results_to_upload_json_trie): Convert a TestRunResults object to a trie to be
uploaded to a results database.
* Scripts/webkitpy/layout_tests/run_webkit_tests.py:
(parse_args): Add new 'Upload Options' and prepare 'Results JSON Options' for deprecation.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243842 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago-apple-trailing-word is needed for browser detection
mmaxfield@apple.com [Thu, 4 Apr 2019 00:16:24 +0000 (00:16 +0000)]
-apple-trailing-word is needed for browser detection
https://bugs.webkit.org/show_bug.cgi?id=196575

Unreviewed.

PerformanceTests:

* MotionMark/resources/debug-runner/motionmark.css:
(#intro .start-benchmark p):

Source/JavaScriptCore:

* Configurations/FeatureDefines.xcconfig:

Source/WebCore:

This is an unreviewed partial revert of r243819. Turns out there are some websites
which use this property to do browser detection. So, we need to continue to parse
the property, but we don't need the property to do anything.

Test: fast/text/trailing-word-detection.html

* Configurations/FeatureDefines.xcconfig:
* css/CSSComputedStyleDeclaration.cpp:
(WebCore::ComputedStyleExtractor::valueForPropertyinStyle):
* css/CSSPrimitiveValueMappings.h:
(WebCore::CSSPrimitiveValue::CSSPrimitiveValue):
(WebCore::CSSPrimitiveValue::operator TrailingWord const):
* css/CSSProperties.json:
* css/CSSValueKeywords.in:
* css/parser/CSSParserFastPaths.cpp:
(WebCore::CSSParserFastPaths::isValidKeywordPropertyAndValue):
(WebCore::CSSParserFastPaths::isKeywordPropertyID):
* rendering/style/RenderStyle.h:
(WebCore::RenderStyle::trailingWord const):
(WebCore::RenderStyle::setTrailingWord):
(WebCore::RenderStyle::initialTrailingWord):
* rendering/style/RenderStyleConstants.h:

Source/WebCore/PAL:

* Configurations/FeatureDefines.xcconfig:

Source/WebKit:

* Configurations/FeatureDefines.xcconfig:

Source/WebKitLegacy/mac:

* Configurations/FeatureDefines.xcconfig:

Tools:

* TestWebKitAPI/Configurations/FeatureDefines.xcconfig:

LayoutTests:

* fast/text/trailing-word-detection-expected.txt: Added.
* fast/text/trailing-word-detection.html: Added.
* platform/gtk/TestExpectations:
* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243841 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agotiled-drawing/scrolling/scroll-snap/scroll-snap-proximity-mainframe.html is a flaky...
sroberts@apple.com [Wed, 3 Apr 2019 23:57:51 +0000 (23:57 +0000)]
tiled-drawing/scrolling/scroll-snap/scroll-snap-proximity-mainframe.html is a flaky failure and timeout
https://bugs.webkit.org/show_bug.cgi?id=196115

Unreviewed test gardening.

* platform/mac-wk2/TestExpectations: Updating test expectations for flaky test

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243840 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoREGRESSION (r243642): com.apple.JavaScriptCore crash in JSC::RegExpObject::execInline
msaboff@apple.com [Wed, 3 Apr 2019 23:51:12 +0000 (23:51 +0000)]
REGRESSION (r243642): com.apple.JavaScriptCore crash in JSC::RegExpObject::execInline
https://bugs.webkit.org/show_bug.cgi?id=196477

Reviewed by Keith Miller.

Source/JavaScriptCore:

The problem here is that when we advance the index by 2 for a character class that only
has non-BMP characters, we might go past the end of the string.  This can happen for
greedy counted character classes that are part of a alternative where there is one
character to match after the greedy non-BMP character class.

The "do we have string left to match" check at the top of the JIT loop for the counted
character class checks to see if index is not equal to the string length.  For non-BMP
character classes, we need to check to see if there are at least 2 characters left.
Therefore we now temporarily add 1 to the current index before comparing.  This checks
to see if there are iat least 2 characters left to match, instead of 1.

* yarr/YarrJIT.cpp:
(JSC::Yarr::YarrGenerator::generateCharacterClassGreedy):
(JSC::Yarr::YarrGenerator::backtrackCharacterClassNonGreedy):

LayoutTests:

Updated the test with a couple more test cases to test a few variants of this bug.
Also added a couple of non-greedy counted non-BMP character class tests that don't have
the bug just to be sure.

* js/regexp-unicode-expected.txt:
* js/script-tests/regexp-unicode.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243839 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoRemove unneeded libwebrtc files, this time all the files
youenn@apple.com [Wed, 3 Apr 2019 23:39:12 +0000 (23:39 +0000)]
Remove unneeded libwebrtc files, this time all the files
https://bugs.webkit.org/show_bug.cgi?id=196553

Reviewed by Eric Carlson.

* Source/third_party/boringssl/src/fuzz: Removed.
* Source/third_party/protobuf/csharp/keys: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243838 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoRemove unneeded libwebrtc files
youenn@apple.com [Wed, 3 Apr 2019 23:35:05 +0000 (23:35 +0000)]
Remove unneeded libwebrtc files
https://bugs.webkit.org/show_bug.cgi?id=196553

Reviewed by Eric Carlson.

* Source/third_party/boringssl/src/fuzz: Removed.
* Source/third_party/protobuf/csharp/keys: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243837 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoAdd a WebKit internal feature flag to always enable modern compatibility mode by...
wenson_hsieh@apple.com [Wed, 3 Apr 2019 23:30:15 +0000 (23:30 +0000)]
Add a WebKit internal feature flag to always enable modern compatibility mode by default
https://bugs.webkit.org/show_bug.cgi?id=196526
<rdar://problem/49532923>

Reviewed by Tim Horton.

Change UseModernCompatibilityModeByDefault from an internal preference to a debug preference, so that it may be
enabled for all apps on the system.

* Shared/WebPreferences.yaml:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243836 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[JSC] Exception verification crash on operationArrayIndexOfValueInt32OrContiguous
ysuzuki@apple.com [Wed, 3 Apr 2019 23:29:48 +0000 (23:29 +0000)]
[JSC] Exception verification crash on operationArrayIndexOfValueInt32OrContiguous
https://bugs.webkit.org/show_bug.cgi?id=196574

Reviewed by Saam Barati.

JSTests:

* stress/string-index-of-exception-check.js: Added.
(blurType):
(1.forEach):

Source/JavaScriptCore:

This patch adds missing exception check in operationArrayIndexOfValueInt32OrContiguous.

* dfg/DFGOperations.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243835 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoFix internal build after r243829
achristensen@apple.com [Wed, 3 Apr 2019 23:21:04 +0000 (23:21 +0000)]
Fix internal build after r243829
https://bugs.webkit.org/show_bug.cgi?id=196549

* NetworkProcess/cocoa/NetworkSessionCocoa.mm:
(WebKit::NetworkSessionCocoa::NetworkSessionCocoa):
Some SDK's don't include this selector in the header.
Use NSSelectorFromString.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243834 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[CMake][WTF] Mirror XCode header directories
don.olmstead@sony.com [Wed, 3 Apr 2019 23:08:44 +0000 (23:08 +0000)]
[CMake][WTF] Mirror XCode header directories
https://bugs.webkit.org/show_bug.cgi?id=191662

Reviewed by Konstantin Tokarev.

.:

Add WTF_FRAMEWORK_HEADERS_DIR to place WTF headers into. Add frameworks/WTF.cmake
which creates an interface target which will populate dependencies for consumers.
This file is added here to support AppleWin internal builds which invoke CMake on
each directory.

* Source/cmake/WebKitFS.cmake:
* Source/cmake/frameworks/WTF.cmake: Added.

Source/JavaScriptCore:

Use WTFFramework as a dependency and include frameworks/WTF.cmake for AppleWin internal
builds.

* CMakeLists.txt:
* shell/CMakeLists.txt:

Source/WebCore:

Use WTFFramework as a dependency and include frameworks/WTF.cmake for AppleWin internal
builds.

* CMakeLists.txt:
* PlatformWin.cmake:

Source/WebDriver:

Use WTFFramework as a dependency.

* CMakeLists.txt:

Source/WebKit:

Use WTFFramework as a dependency and include frameworks/WTF.cmake for AppleWin internal
builds.

* CMakeLists.txt:

Source/WebKitLegacy:

Use WTFFramework as a dependency and include frameworks/WTF.cmake for AppleWin internal
builds.

* CMakeLists.txt:
* PlatformWin.cmake:

Source/WebKitLegacy/win:

* WebKitQuartzCoreAdditions/CMakeLists.txt:

Source/WTF:

Rename WTF forwarding header target to WTFFramework and update the install location
to WTF_FRAMEWORK_HEADERS_DIR.

* wtf/CMakeLists.txt:

Tools:

Use WTFFramework as a dependency and include frameworks/WTF.cmake for AppleWin internal
builds.

* CMakeLists.txt:
* DumpRenderTree/CMakeLists.txt:
* DumpRenderTree/PlatformWin.cmake:
* MiniBrowser/win/CMakeLists.txt:
* TestWebKitAPI/CMakeLists.txt:
* TestWebKitAPI/PlatformWin.cmake:
* WebKitTestRunner/PlatformGTK.cmake:
* WebKitTestRunner/PlatformWin.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243833 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[JSC] Add FuzzerAgent, which has a hooks to get feedback & inject fuzz data into JSC
ysuzuki@apple.com [Wed, 3 Apr 2019 22:24:47 +0000 (22:24 +0000)]
[JSC] Add FuzzerAgent, which has a hooks to get feedback & inject fuzz data into JSC
https://bugs.webkit.org/show_bug.cgi?id=196530

Reviewed by Saam Barati.

This patch adds FuzzerAgent interface and simple RandomizingFuzzerAgent to JSC.
This RandomizingFuzzerAgent returns random SpeculatedType for value profiling to find
the issues in JSC. The seed for randomization can be specified by seedOfRandomizingFuzzerAgent.

I ran this with seedOfRandomizingFuzzerAgent=1 last night and it finds 3 failures in the current JSC tests,
they should be fixed in subsequent patches.

* CMakeLists.txt:
* JavaScriptCore.xcodeproj/project.pbxproj:
* Sources.txt:
* dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::getPredictionWithoutOSRExit):
* runtime/FuzzerAgent.cpp: Added.
(JSC::FuzzerAgent::~FuzzerAgent):
(JSC::FuzzerAgent::getPrediction):
* runtime/FuzzerAgent.h: Added.
* runtime/JSGlobalObjectFunctions.cpp:
* runtime/Options.h:
* runtime/RandomizingFuzzerAgent.cpp: Added.
(JSC::RandomizingFuzzerAgent::RandomizingFuzzerAgent):
(JSC::RandomizingFuzzerAgent::getPrediction):
* runtime/RandomizingFuzzerAgent.h: Added.
* runtime/RegExpCachedResult.h:
* runtime/RegExpGlobalData.cpp:
* runtime/VM.cpp:
(JSC::VM::VM):
* runtime/VM.h:
(JSC::VM::fuzzerAgent const):
(JSC::VM::setFuzzerAgent):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243832 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoAvoid the use of WTF_EXTERN_C_BEGIN in NSMenuSPI.h
rniwa@webkit.org [Wed, 3 Apr 2019 22:16:57 +0000 (22:16 +0000)]
Avoid the use of WTF_EXTERN_C_BEGIN in NSMenuSPI.h
https://bugs.webkit.org/show_bug.cgi?id=196566

Reviewed by Timothy Hatcher.

* pal/spi/mac/NSMenuSPI.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243831 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoRemove SVG properties tear-off objects
commit-queue@webkit.org [Wed, 3 Apr 2019 22:07:59 +0000 (22:07 +0000)]
Remove SVG properties tear-off objects
https://bugs.webkit.org/show_bug.cgi?id=191237

Patch by Said Abou-Hallawa <sabouhallawa@apple.com> on 2019-04-03
Reviewed by Simon Fraser.

This patch cleans the code from unused SVG sources and get rid off the
remaining SVG properties tear-off objects. Here are more details:

-- Remove the SVGAttributeAnimationController and move its code to the
   SVGAnimateElementBase. SVGAttributeAnimationController was introduced
   to allow animating the SVG properties whether they are backed by tear
   off objects or not. Since there will be no tear off objects anymore,
   one animation controller will be needed. But in this case, it will be
   better if we make SVGAnimateElementBase is the animation controller
   and make it manage the animator directly.

-- Remove SVGAttributeRegistry, SVGAttributeOwnerProxy and the virtual
   function SVGElement::attributeOwnerProxy(). Remove also all the
   overriding functions attributeOwnerProxy() from all the SVGElements.

-- Remove isKnownAttribute() from all the SVG header files except from
   four classes: SVGURIReference, SVGFitToViewBox, SVGLangSpace
   and SVGExternalResourcesRequired.

-- Remove all the SVG animated properties classifying functions from
   SVGElement. This is now handled by SVGPropertyRegistry.

-- There is no need for the enum AnimatedPropertyType anymore. The SVG
   property accessor knows its type, knows how to access it and know what
   animator it should be created for it.

* Sources.txt:
* WebCore.xcodeproj/project.pbxproj:
* dom/Element.cpp:
(WebCore::Element::synchronizeAllAttributes const):
(WebCore::Element::synchronizeAttribute const):
(WebCore::Element::fastAttributeLookupAllowed const):
These functions are surprisingly marked 'const'. They were calling 'const'
functions in SVGElement and SVGElement was casting 'this' as non 'const'
before calling the non 'const' functions through the non 'const' 'this'
pointer. Change this by moving the casting to the Element functions.

* rendering/svg/RenderSVGResourceGradient.cpp:
(WebCore::RenderSVGResourceGradient::applyResource):
* rendering/svg/RenderSVGResourcePattern.cpp:
(WebCore::RenderSVGResourcePattern::applyResource):
* svg/SVGAElement.h:
* svg/SVGAltGlyphElement.h:
* svg/SVGAnimateElementBase.cpp:
(WebCore::SVGAnimateElementBase::SVGAnimateElementBase):
(WebCore::SVGAnimateElementBase::animator const):
(WebCore::SVGAnimateElementBase::hasInvalidCSSAttributeType const):
(WebCore::SVGAnimateElementBase::isDiscreteAnimator const):
(WebCore::SVGAnimateElementBase::setTargetElement):
(WebCore::SVGAnimateElementBase::setAttributeName):
(WebCore::SVGAnimateElementBase::resetAnimation):
(WebCore::SVGAnimateElementBase::calculateFromAndToValues):
(WebCore::SVGAnimateElementBase::calculateFromAndByValues):
(WebCore::SVGAnimateElementBase::calculateToAtEndOfDurationValue):
(WebCore::SVGAnimateElementBase::resetAnimatedType):
(WebCore::SVGAnimateElementBase::calculateAnimatedValue):
(WebCore::SVGAnimateElementBase::applyResultsToTarget):
(WebCore::SVGAnimateElementBase::clearAnimatedType):
(WebCore::SVGAnimateElementBase::calculateDistance):
(WebCore::SVGAnimateElementBase::attributeAnimationController): Deleted.
(WebCore::SVGAnimateElementBase::determineAnimatedPropertyType const): Deleted.
(WebCore::SVGAnimateElementBase::isAdditive const): Deleted.
* svg/SVGAnimateElementBase.h:
(WebCore::SVGAnimateElementBase::animatorIfExists const):
(WebCore::SVGAnimateElementBase::attributeAnimationControllerIfExists const): Deleted.
The order of the functions was changed to resemble the order of running
the animation:
  -- Setting the animation range
  -- Starting the animation
  -- Progressing the animation
  -- Applying the the progressed animaVal() to the target element
  -- Stopping the animation

* svg/SVGAnimateMotionElement.cpp:
(WebCore::SVGAnimateMotionElement::calculateDistance):
* svg/SVGAnimateMotionElement.h:
* svg/SVGAnimationElement.cpp:
(WebCore::SVGAnimationElement::calculateKeyTimesForCalcModePaced):
(WebCore::SVGAnimationElement::shouldApplyAnimation): Deleted.
* svg/SVGAnimationElement.h:
(WebCore::SVGAnimationElement::attributeRegistry): Deleted.
(WebCore::SVGAnimationElement::calculateDistance): Deleted.
* svg/SVGAttributeAnimationController.cpp: Removed.
* svg/SVGAttributeAnimationController.h: Removed.
* svg/SVGAttributeAnimationControllerBase.cpp: Removed.
* svg/SVGAttributeAnimationControllerBase.h: Removed.
* svg/SVGCircleElement.h:
* svg/SVGClipPathElement.h:
* svg/SVGComponentTransferFunctionElement.h:
* svg/SVGCursorElement.h:
* svg/SVGDefsElement.h:
* svg/SVGElement.cpp:
(WebCore::SVGElement::synchronizeAttribute):
(WebCore::SVGElement::synchronizeAllAttributes):
(WebCore::SVGElement::synchronizeAllAnimatedSVGAttribute):
(WebCore::createAttributeNameToAnimatedPropertyTypeMap): Deleted.
(WebCore::attributeNameToAnimatedPropertyTypeMap): Deleted.
(WebCore::createCSSPropertyWithSVGDOMNameToAnimatedPropertyTypeMap): Deleted.
(WebCore::cssPropertyWithSVGDOMNameToAnimatedPropertyTypeMap): Deleted.
(WebCore::SVGElement::animatedPropertyTypesForAttribute): Deleted.
(WebCore::SVGElement::synchronizeAnimatedSVGAttribute const): Deleted.
(WebCore::SVGElement::isAnimatableCSSProperty): Deleted.
(WebCore::SVGElement::isPresentationAttributeWithSVGDOM): Deleted.
* svg/SVGElement.h:
(WebCore::SVGElement::attributeOwnerProxy const): Deleted.
(WebCore::SVGElement::attributeRegistry): Deleted.
(WebCore::SVGElement::synchronizeAttribute): Deleted.
(WebCore::SVGElement::synchronizeAttributes): Deleted.
(WebCore::SVGElement::animatedTypes const): Deleted.
(WebCore::SVGElement::lookupAnimatedProperty const): Deleted.
(WebCore::SVGElement::lookupOrCreateAnimatedProperty): Deleted.
(WebCore::SVGElement::lookupOrCreateAnimatedProperties): Deleted.
(WebCore::SVGElement::isKnownAttribute): Deleted.
* svg/SVGEllipseElement.h:
* svg/SVGExternalResourcesRequired.h:
(WebCore::SVGExternalResourcesRequired::attributeRegistry): Deleted.
(WebCore::SVGExternalResourcesRequired::attributeOwnerProxy): Deleted.
* svg/SVGFEBlendElement.h:
* svg/SVGFEColorMatrixElement.h:
* svg/SVGFEComponentTransferElement.h:
* svg/SVGFECompositeElement.h:
* svg/SVGFEConvolveMatrixElement.h:
* svg/SVGFEDiffuseLightingElement.h:
* svg/SVGFEDisplacementMapElement.h:
* svg/SVGFEDropShadowElement.h:
* svg/SVGFEGaussianBlurElement.h:
* svg/SVGFEImageElement.h:
* svg/SVGFELightElement.h:
* svg/SVGFEMergeNodeElement.h:
* svg/SVGFEMorphologyElement.h:
* svg/SVGFEOffsetElement.h:
* svg/SVGFESpecularLightingElement.h:
* svg/SVGFETileElement.h:
* svg/SVGFETurbulenceElement.h:
* svg/SVGFilterElement.h:
* svg/SVGFilterPrimitiveStandardAttributes.h:
(WebCore::SVGFilterPrimitiveStandardAttributes::attributeRegistry): Deleted.
* svg/SVGFitToViewBox.h:
(WebCore::SVGFitToViewBox::attributeRegistry): Deleted.
* svg/SVGFontElement.h:
* svg/SVGForeignObjectElement.h:
* svg/SVGGElement.h:
* svg/SVGGeometryElement.h:
* svg/SVGGlyphRefElement.h:
* svg/SVGGradientElement.cpp:
(WebCore::SVGGradientElement::svgAttributeChanged):
* svg/SVGGradientElement.h:
(WebCore::SVGGradientElement::attributeRegistry): Deleted.
(WebCore::SVGGradientElement::isKnownAttribute): Deleted.
* svg/SVGGraphicsElement.h:
(WebCore::SVGGraphicsElement::attributeRegistry): Deleted.
* svg/SVGImageElement.h:
* svg/SVGLineElement.h:
* svg/SVGLinearGradientElement.h:
* svg/SVGMPathElement.h:
* svg/SVGMarkerElement.cpp:
* svg/SVGMarkerElement.h:
* svg/SVGMaskElement.h:
* svg/SVGPathElement.h:
* svg/SVGPatternElement.h:
* svg/SVGPolyElement.h:
* svg/SVGRadialGradientElement.h:
* svg/SVGRectElement.h:
* svg/SVGSVGElement.h:
* svg/SVGScriptElement.h:
* svg/SVGStopElement.h:
* svg/SVGSwitchElement.h:
* svg/SVGSymbolElement.h:
* svg/SVGTRefElement.h:
* svg/SVGTests.cpp:
(WebCore::SVGTests::svgAttributeChanged):
(WebCore::SVGTests::attributeRegistry): Deleted.
(WebCore::SVGTests::isKnownAttribute): Deleted.
* svg/SVGTests.h:
* svg/SVGTextContentElement.cpp:
* svg/SVGTextContentElement.h:
(WebCore::SVGTextContentElement::attributeRegistry): Deleted.
* svg/SVGTextPathElement.h:
* svg/SVGTextPositioningElement.h:
(WebCore::SVGTextPositioningElement::attributeRegistry): Deleted.
* svg/SVGURIReference.cpp:
(WebCore::SVGURIReference::SVGURIReference):
(WebCore::SVGURIReference::attributeRegistry): Deleted.
* svg/SVGURIReference.h:
* svg/SVGUseElement.h:
* svg/SVGViewElement.h:
* svg/SVGViewSpec.cpp:
(WebCore::SVGViewSpec::SVGViewSpec):
* svg/SVGViewSpec.h:
* svg/SVGZoomAndPanType.h:
* svg/properties/SVGAnimatedListPropertyTearOff.h: Removed.
* svg/properties/SVGAnimatedPropertyAnimator.h:
* svg/properties/SVGAnimatedPropertyAnimatorImpl.h:
* svg/properties/SVGAnimatedPropertyPairAnimator.h:
* svg/properties/SVGAnimatedPropertyPairAnimatorImpl.h:
* svg/properties/SVGAnimatedPropertyTearOff.h: Removed.
* svg/properties/SVGAnimatedPropertyType.h: Removed.

* svg/properties/SVGAnimationAdditiveFunction.h:
(WebCore::SVGAnimationAdditiveFunction::animate):
(WebCore::SVGAnimationAdditiveFunction::progress): Deleted.
* svg/properties/SVGAnimationAdditiveListFunctionImpl.h:
(WebCore::SVGAnimationLengthListFunction::animate):
(WebCore::SVGAnimationNumberListFunction::animate):
(WebCore::SVGAnimationPointListFunction::animate):
(WebCore::SVGAnimationTransformListFunction::animate):
(WebCore::SVGAnimationLengthListFunction::progress): Deleted.
(WebCore::SVGAnimationNumberListFunction::progress): Deleted.
(WebCore::SVGAnimationPointListFunction::progress): Deleted.
(WebCore::SVGAnimationTransformListFunction::progress): Deleted.
* svg/properties/SVGAnimationAdditiveValueFunctionImpl.h:
(WebCore::SVGAnimationAngleFunction::animate):
(WebCore::SVGAnimationColorFunction::animate):
(WebCore::SVGAnimationIntegerFunction::animate):
(WebCore::SVGAnimationLengthFunction::animate):
(WebCore::SVGAnimationNumberFunction::animate):
(WebCore::SVGAnimationPathSegListFunction::animate):
(WebCore::SVGAnimationRectFunction::animate):
(WebCore::SVGAnimationAngleFunction::progress): Deleted.
(WebCore::SVGAnimationColorFunction::progress): Deleted.
(WebCore::SVGAnimationIntegerFunction::progress): Deleted.
(WebCore::SVGAnimationLengthFunction::progress): Deleted.
(WebCore::SVGAnimationNumberFunction::progress): Deleted.
(WebCore::SVGAnimationPathSegListFunction::progress): Deleted.
(WebCore::SVGAnimationRectFunction::progress): Deleted.
* svg/properties/SVGAnimationDiscreteFunction.h:
(WebCore::SVGAnimationDiscreteFunction::animate):
(WebCore::SVGAnimationDiscreteFunction::progress): Deleted.
* svg/properties/SVGAnimationFunction.h:
(WebCore::SVGAnimationFunction::calculateDistance const):
-- Rename the 'progress()' functions of SVGAttributeAnimator and SVGAnimationFunction
   to 'animate()'.
-- Rename the argument 'percentage' of these function to 'progress'
-- Make calculateDistance return Optional<float> so it does not have to
   return -1 in case of error.

* svg/properties/SVGAttribute.h: Removed.
* svg/properties/SVGAttributeAccessor.h: Removed.
* svg/properties/SVGAttributeAnimator.h:
(WebCore::SVGAttributeAnimator::calculateDistance const):
* svg/properties/SVGAttributeOwnerProxy.cpp: Removed.
* svg/properties/SVGAttributeOwnerProxy.h: Removed.
* svg/properties/SVGAttributeOwnerProxyImpl.h: Removed.
* svg/properties/SVGAttributeRegistry.h: Removed.
* svg/properties/SVGLegacyAnimatedProperty.cpp: Removed.
* svg/properties/SVGLegacyAnimatedProperty.h: Removed.
* svg/properties/SVGLegacyProperty.h: Removed.
* svg/properties/SVGListProperty.h: Removed.
* svg/properties/SVGListPropertyTearOff.h: Removed.
* svg/properties/SVGPrimitivePropertyAnimator.h:
* svg/properties/SVGProperty.h:
* svg/properties/SVGPropertyAnimator.h:
* svg/properties/SVGPropertyTearOff.h: Removed.
* svg/properties/SVGValuePropertyAnimator.h:
* svg/properties/SVGValuePropertyListAnimator.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243830 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoAdd SPI to disable legacy TLS fallback
commit-queue@webkit.org [Wed, 3 Apr 2019 22:04:57 +0000 (22:04 +0000)]
Add SPI to disable legacy TLS fallback
https://bugs.webkit.org/show_bug.cgi?id=196549
<rdar://44979744>

Patch by Alex Christensen <achristensen@webkit.org> on 2019-04-03
Reviewed by Geoffrey Garen.

Source/WebCore/PAL:

* pal/spi/cf/CFNetworkSPI.h:

Source/WebKit:

* NetworkProcess/NetworkSessionCreationParameters.cpp:
(WebKit::NetworkSessionCreationParameters::privateSessionParameters):
(WebKit::NetworkSessionCreationParameters::encode const):
(WebKit::NetworkSessionCreationParameters::decode):
* NetworkProcess/NetworkSessionCreationParameters.h:
* NetworkProcess/cocoa/NetworkSessionCocoa.mm:
(WebKit::NetworkSessionCocoa::NetworkSessionCocoa):
* UIProcess/API/Cocoa/WKWebsiteDataStore.mm:
(-[WKWebsiteDataStore _setAllowsTLSFallback:]):
(-[WKWebsiteDataStore _allowsTLSFallback]):
* UIProcess/API/Cocoa/WKWebsiteDataStorePrivate.h:
* UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::ensureNetworkProcess):
* UIProcess/WebsiteData/Cocoa/WebsiteDataStoreCocoa.mm:
(WebKit::WebsiteDataStore::parameters):
* UIProcess/WebsiteData/WebsiteDataStore.cpp:
(WebKit::WebsiteDataStore::setSourceApplicationSecondaryIdentifier):
(WebKit::WebsiteDataStore::setAllowsTLSFallback):
(WebKit::WebsiteDataStore::setSourceApplicationBundleIdentifier):
* UIProcess/WebsiteData/WebsiteDataStore.h:
(WebKit::WebsiteDataStore::allowsTLSFallback const):
(WebKit::WebsiteDataStore::networkingHasBegun):
(WebKit::WebsiteDataStore::finalizeApplicationIdentifiers): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243829 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoDocuments can be destroyed before their CSSFontFaceSet is destroyed
mmaxfield@apple.com [Wed, 3 Apr 2019 21:46:55 +0000 (21:46 +0000)]
Documents can be destroyed before their CSSFontFaceSet is destroyed
https://bugs.webkit.org/show_bug.cgi?id=195830

Reviewed by Darin Adler.

Source/WebCore:

CSSFontFaceSet has a raw pointer to its owning document. JS can keep the CSSFontFaceSet alive (by using FontFaceSet)
and can destroy the document at any time. When the document is destroyed, the link between the two objects needs to
be severed.

Test: fast/text/font-face-set-destroy-document.html

* css/CSSFontFace.cpp:
(WebCore::CSSFontFace::CSSFontFace):
* css/CSSFontFace.h:
* css/CSSFontFaceSet.cpp:
(WebCore::CSSFontFaceSet::CSSFontFaceSet):
(WebCore::CSSFontFaceSet::ensureLocalFontFacesForFamilyRegistered):
* css/CSSFontFaceSet.h:
* css/CSSFontSelector.cpp:
(WebCore::CSSFontSelector::CSSFontSelector):
(WebCore::CSSFontSelector::addFontFaceRule):
* css/CSSFontSelector.h:
* css/FontFace.cpp:
(WebCore::FontFace::FontFace):

LayoutTests:

* fast/text/font-face-set-destroy-document-expected.html: Added.
* fast/text/font-face-set-destroy-document.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243828 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agohttp/tests/storageAccess/request-and-grant-access-cross-origin-sandboxed-iframe-from...
sroberts@apple.com [Wed, 3 Apr 2019 21:29:23 +0000 (21:29 +0000)]
http/tests/storageAccess/request-and-grant-access-cross-origin-sandboxed-iframe-from-prevalent-domain-with-user-interaction-but-access-from-wrong-frame.html is a flaky timeout
https://bugs.webkit.org/show_bug.cgi?id=194164

Unreviewed test gardening.

* platform/mac-wk2/TestExpectations: Updating test expectations for flaky timeout

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243827 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoWeb Inspector: Single click on links in non-read-only TextEditors should not follow...
drousso@apple.com [Wed, 3 Apr 2019 21:03:58 +0000 (21:03 +0000)]
Web Inspector: Single click on links in non-read-only TextEditors should not follow links
https://bugs.webkit.org/show_bug.cgi?id=123364
<rdar://problem/15323913>

Reviewed by Timothy Hatcher.

* UserInterface/Base/Main.js:
(WI._updateModifierKeys):
Add classes to the body whenever alt, shift, or ctrl are pressed.

* UserInterface/Views/CodeMirrorEditor.js:
(WI.CodeMirrorEditor.create):
Add a `read-only` class if the `CodeMirror` is readonly.

* UserInterface/Views/SyntaxHighlightingDefaultTheme.css:
(.cm-s-default .cm-link,):
(.read-only.cm-s-default .cm-link:hover,):
(.cm-s-default .cm-link:hover,): Deleted.

* UserInterface/Views/TextEditor.js:
(WI.TextEditor.prototype.set readOnly):
(WI.TextEditor.prototype._openClickedLinks):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243826 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoRebaseline scrollingcoordinator/scrolling-tree/fixed-inside-frame.html
sroberts@apple.com [Wed, 3 Apr 2019 20:49:17 +0000 (20:49 +0000)]
Rebaseline scrollingcoordinator/scrolling-tree/fixed-inside-frame.html
https://bugs.webkit.org/show_bug.cgi?id=195254

Unreviewed test gardening.

* scrollingcoordinator/scrolling-tree/fixed-inside-frame-expected.txt: Rebaselined test

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243825 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoFollow up fix for r243807: Use MarkedArgumentBuffer instead of Vector for JSValue
sihui_liu@apple.com [Wed, 3 Apr 2019 20:37:22 +0000 (20:37 +0000)]
Follow up fix for r243807: Use MarkedArgumentBuffer instead of Vector for JSValue
https://bugs.webkit.org/show_bug.cgi?id=196547

Reviewed by Geoffrey Garen.

JSValue in Vector could be garbage collected because GC doesn't know Vector memory on C++ heap.

* bindings/js/JSIDBRequestCustom.cpp:
(WebCore::JSIDBRequest::result const):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243824 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[ews-app] Status bubble should not display certain infrastructure status messages
aakash_jain@apple.com [Wed, 3 Apr 2019 20:27:12 +0000 (20:27 +0000)]
[ews-app] Status bubble should not display certain infrastructure status messages
https://bugs.webkit.org/show_bug.cgi?id=196550

Reviewed by Lucas Forschler.

Do not display the steps which are un-needed for end-user. This information is still avaiable
if the user click on the bubble and look at the Buildbot build.
* BuildSlaveSupport/ews-app/ews/views/statusbubble.py:
(StatusBubble._steps_messages):
(StatusBubble._should_display_step): Method to decide whether to display the step or not.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243823 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[ews-build] Use https for sending data to ews-app
aakash_jain@apple.com [Wed, 3 Apr 2019 20:13:42 +0000 (20:13 +0000)]
[ews-build] Use https for sending data to ews-app
https://bugs.webkit.org/show_bug.cgi?id=196492

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-build/events.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243822 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoHTML fragment serialization should not strip whitespace from URL attribute values
cdumez@apple.com [Wed, 3 Apr 2019 20:08:34 +0000 (20:08 +0000)]
HTML fragment serialization should not strip whitespace from URL attribute values
https://bugs.webkit.org/show_bug.cgi?id=196551

Reviewed by Ryosuke Niwa.

LayoutTests/imported/w3c:

Rebaseline WPT test now that all checks are passing. This test was already passing
in Gecko and Blink.

* web-platform-tests/domparsing/innerhtml-mxss.sub-expected.txt:

Source/WebCore:

HTML fragment serialization should not strip whitespace from URL attribute values as per:
- https://html.spec.whatwg.org/multipage/parsing.html#html-fragment-serialisation-algorithm

WebKit was stripping such whitespace, Gecko and Blink are not. Align WebKit with other
browser engines and the specification.

No new tests, rebaselined existing test.

* editing/MarkupAccumulator.cpp:
(WebCore::MarkupAccumulator::appendQuotedURLAttributeValue):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243821 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoCrash in HTMLCanvasElement::createContext2d after the element got adopted to a new...
rniwa@webkit.org [Wed, 3 Apr 2019 20:04:37 +0000 (20:04 +0000)]
Crash in HTMLCanvasElement::createContext2d after the element got adopted to a new document
https://bugs.webkit.org/show_bug.cgi?id=196527

Reviewed by Antti Koivisto.

We need to update CanvasBase::m_scriptExecutionContext when HTMLCanvasElement moves from
one document to another. Fixed the bug by making CanvasBase::scriptExecutionContext make
a virtual function call instead of directly storing a raw pointer. In HTMLCanvasElement,
we use Node::scriptExecutionContext(). Use ContextDestructionObserver in CustomPaintCanvas
and OffscreenCanvas instead of a raw pointer.

Unfortunately, no new tests since there is no reproducible test case.

* html/CanvasBase.cpp:
(WebCore::CanvasBase::CanvasBase):
* html/CanvasBase.h:
(WebCore::CanvasBase::scriptExecutionContext const):
* html/CustomPaintCanvas.cpp:
(WebCore::CustomPaintCanvas::CustomPaintCanvas):
* html/CustomPaintCanvas.h:
* html/HTMLCanvasElement.cpp:
(WebCore::HTMLCanvasElement::HTMLCanvasElement):
* html/HTMLCanvasElement.h:
* html/OffscreenCanvas.cpp:
(WebCore::OffscreenCanvas::OffscreenCanvas):
* html/OffscreenCanvas.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243820 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoRemove support for -apple-trailing-word
mmaxfield@apple.com [Wed, 3 Apr 2019 19:47:58 +0000 (19:47 +0000)]
Remove support for -apple-trailing-word
https://bugs.webkit.org/show_bug.cgi?id=196525

Reviewed by Zalan Bujtas.

This CSS property is nonstandard and not used.

.:

* Source/cmake/WebKitFeatures.cmake:

Source/JavaScriptCore:

* Configurations/FeatureDefines.xcconfig:

Source/WebCore:

* Configurations/FeatureDefines.xcconfig:
* css/CSSComputedStyleDeclaration.cpp:
(WebCore::ComputedStyleExtractor::valueForPropertyinStyle):
* css/CSSPrimitiveValueMappings.h:
(WebCore::CSSPrimitiveValue::operator TrailingWord const): Deleted.
* css/CSSProperties.json:
* css/CSSValueKeywords.in:
* css/parser/CSSParserFastPaths.cpp:
(WebCore::CSSParserFastPaths::isValidKeywordPropertyAndValue):
(WebCore::CSSParserFastPaths::isKeywordPropertyID):
* rendering/SimpleLineLayout.cpp:
(WebCore::SimpleLineLayout::canUseForStyle):
* rendering/SimpleLineLayoutCoverage.cpp:
(WebCore::SimpleLineLayout::printReason):
* rendering/SimpleLineLayoutCoverage.h:
* rendering/line/BreakingContext.h:
(WebCore::BreakingContext::BreakingContext):
(WebCore::BreakingContext::lineBreak):
(WebCore::BreakingContext::clearLineBreakIfFitsOnLine):
(WebCore::BreakingContext::commitLineBreakClear):
(WebCore::BreakingContext::commitLineBreakAtCurrentWidth):
(WebCore::BreakingContext::handleBR):
(WebCore::BreakingContext::handleFloat):
(WebCore::BreakingContext::handleText):
(WebCore::BreakingContext::handleEndOfLine):
(WebCore::BreakingContext::InlineIteratorHistory::InlineIteratorHistory): Deleted.
(WebCore::BreakingContext::InlineIteratorHistory::push): Deleted.
(WebCore::BreakingContext::InlineIteratorHistory::update): Deleted.
(WebCore::BreakingContext::InlineIteratorHistory::renderer const): Deleted.
(WebCore::BreakingContext::InlineIteratorHistory::offset const): Deleted.
(WebCore::BreakingContext::InlineIteratorHistory::nextBreakablePosition const): Deleted.
(WebCore::BreakingContext::InlineIteratorHistory::atTextParagraphSeparator const): Deleted.
(WebCore::BreakingContext::InlineIteratorHistory::previousInSameNode const): Deleted.
(WebCore::BreakingContext::InlineIteratorHistory::get const): Deleted.
(WebCore::BreakingContext::InlineIteratorHistory::current const): Deleted.
(WebCore::BreakingContext::InlineIteratorHistory::historyLength const): Deleted.
(WebCore::BreakingContext::InlineIteratorHistory::moveTo): Deleted.
(WebCore::BreakingContext::InlineIteratorHistory::increment): Deleted.
(WebCore::BreakingContext::InlineIteratorHistory::clear): Deleted.
(WebCore::BreakingContext::optimalLineBreakLocationForTrailingWord): Deleted.
* rendering/style/RenderStyle.h:
(WebCore::RenderStyle::trailingWord const): Deleted.
(WebCore::RenderStyle::setTrailingWord): Deleted.
(WebCore::RenderStyle::initialTrailingWord): Deleted.
* rendering/style/RenderStyleConstants.h:
* rendering/style/StyleRareInheritedData.cpp:
(WebCore::StyleRareInheritedData::StyleRareInheritedData):
(WebCore::StyleRareInheritedData::operator== const):
* rendering/style/StyleRareInheritedData.h:

Source/WebCore/PAL:

* Configurations/FeatureDefines.xcconfig:

Source/WebInspectorUI:

* UserInterface/Models/CSSKeywordCompletions.js:

Source/WebKit:

* Configurations/FeatureDefines.xcconfig:

Source/WebKitLegacy/mac:

* Configurations/FeatureDefines.xcconfig:

Tools:

* Scripts/webkitperl/FeatureList.pm:
* TestWebKitAPI/Configurations/FeatureDefines.xcconfig:

LayoutTests:

* fast/text/trailing-word-expected.html: Removed.
* fast/text/trailing-word.html: Removed.
* platform/gtk/TestExpectations:
* platform/mac/fast/text/trailing-word-parse-expected.txt: Removed.
* platform/mac/fast/text/trailing-word-parse.html: Removed.
* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243819 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoUse makePendingActivity in DOMCache
youenn@apple.com [Wed, 3 Apr 2019 19:38:51 +0000 (19:38 +0000)]
Use makePendingActivity in DOMCache
https://bugs.webkit.org/show_bug.cgi?id=196515

Reviewed by Geoffrey Garen.

No change of behavior, just modernizing the code.

* Modules/cache/DOMCache.cpp:
(WebCore::DOMCache::retrieveRecords):
(WebCore::DOMCache::batchDeleteOperation):
(WebCore::DOMCache::batchPutOperation):
* Modules/cache/DOMCacheStorage.cpp:
(WebCore::DOMCacheStorage::match):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243818 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[XML Parser] Insert the error message block when stopping parsing and an error occurred
cdumez@apple.com [Wed, 3 Apr 2019 19:13:49 +0000 (19:13 +0000)]
[XML Parser] Insert the error message block when stopping parsing and an error occurred
https://bugs.webkit.org/show_bug.cgi?id=196546

Reviewed by Alexey Proskuryakov.

LayoutTests/imported/w3c:

Rebaseline WPT test now that all checks are passing. Both Gecko and Blink already passed
these checks.

* web-platform-tests/domparsing/DOMParser-parseFromString-xml-parsererror-expected.txt:

Source/WebCore:

Insert the error message block when stopping parsing and an error occurred. This is based
on the following Blink commit:
- https://chromium.googlesource.com/chromium/src.git/+/565958bc22e2d49fed7af144482c2bf4d416fec5

No new tests, rebaselined existing test.

* xml/parser/XMLDocumentParser.cpp:
(WebCore::XMLDocumentParser::end):
Avoid showing the error message block twice in some cases. No need to ever call
insertErrorMessageBlock() if we're already stopped since stopParsing() already
takes care of doing this.

* xml/parser/XMLDocumentParserLibxml2.cpp:
(WebCore::XMLDocumentParser::stopParsing):
When XMLDocumentParser::stopParsing() is called to stop parsing, call
insertErrorMessageBlock() to insert the <parsererror> element if an error
occurred.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243817 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoWeb Inspector: Uncaught Exception: TypeError: this._computedStyleSection.focus is...
nvasilyev@apple.com [Wed, 3 Apr 2019 18:05:03 +0000 (18:05 +0000)]
Web Inspector: Uncaught Exception: TypeError: this._computedStyleSection.focus is not a function. (In 'this._computedStyleSection.focus()', 'this._computedStyleSection.focus' is undefined)
https://bugs.webkit.org/show_bug.cgi?id=192108
<rdar://problem/46685392>

Reviewed by Devin Rousso.

This patch fixes the uncaught exception. Tabbing to and from "Properties" section continues
working the same way as before the patch.

* UserInterface/Views/ComputedStyleDetailsPanel.js:
(WI.ComputedStyleDetailsPanel.prototype.focusFirstSection): Deleted.
(WI.ComputedStyleDetailsPanel.prototype.focusLastSection): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243816 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoAdding myself to contributors.json.
kocsen_chung@apple.com [Wed, 3 Apr 2019 18:03:30 +0000 (18:03 +0000)]
Adding myself to contributors.json.

Unreviewed contributors.json update.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243815 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoAdd a test for DOMCache on multiple browsing in case of private browsing
youenn@apple.com [Wed, 3 Apr 2019 17:54:44 +0000 (17:54 +0000)]
Add a test for DOMCache on multiple browsing in case of private browsing
https://bugs.webkit.org/show_bug.cgi?id=196524

Reviewed by Alex Christensen.

* TestWebKitAPI/Tests/WebKitCocoa/ServiceWorkerBasic.mm:
(-[SWMessageHandlerForCacheStorage userContentController:didReceiveScriptMessage:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243814 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoRe-sync web-platform-tests/domparsing from upstream
cdumez@apple.com [Wed, 3 Apr 2019 17:48:31 +0000 (17:48 +0000)]
Re-sync web-platform-tests/domparsing from upstream
https://bugs.webkit.org/show_bug.cgi?id=196544

Reviewed by Alex Christensen.

Re-sync web-platform-tests/domparsing from upstream 3bfdeb8976fc5.

* web-platform-tests/domparsing/*: Updated
* web-platform-tests/interfaces/DOM-Parsing.idl: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243813 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoClear WorkerCacheStorageConnection callbacks on WorkerGlobalScope termination
youenn@apple.com [Wed, 3 Apr 2019 17:45:17 +0000 (17:45 +0000)]
Clear WorkerCacheStorageConnection callbacks on WorkerGlobalScope termination
https://bugs.webkit.org/show_bug.cgi?id=196521

Reviewed by Alex Christensen.

When the worker global scope is preparing for termination,
all ActiveDOMObjects are stopped.
At that time, the completion handlers related to
WorkerCacheStorageConnection should be cleared to be able to free
memory, and as they are now no-op anyway.

We clear the completion handlers once the active DOM objects are stopped
to limit the processing triggered by clearing them.

Introducing a new Stopped error code to handle this case.
Add an assertion so that this error does not surface to JS.

Covered by existing tests.

* Modules/cache/CacheStorageConnection.cpp:
(WebCore::CacheStorageConnection::clearPendingRequests):
* Modules/cache/CacheStorageConnection.h:
* Modules/cache/DOMCacheEngine.cpp:
(WebCore::DOMCacheEngine::errorToException):
* Modules/cache/DOMCacheEngine.h:
* workers/WorkerGlobalScope.cpp:
(WebCore::WorkerGlobalScope::prepareForTermination):
(WebCore::WorkerGlobalScope::stopIndexedDatabase):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243812 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoAdopt new VCP SPI
youenn@apple.com [Wed, 3 Apr 2019 17:37:55 +0000 (17:37 +0000)]
Adopt new VCP SPI
https://bugs.webkit.org/show_bug.cgi?id=193357
<rdar://problem/43656651>

Reviewed by Eric Carlson.

Source/ThirdParty/libwebrtc:

       Enable VCP through VTB API with specific encoder id.

* Source/webrtc/sdk/WebKit/VideoProcessingSoftLink.cpp:
* Source/webrtc/sdk/WebKit/VideoProcessingSoftLink.h:
* Source/webrtc/sdk/WebKit/WebKitUtilities.mm:
(webrtc::setApplicationStatus):
* Source/webrtc/sdk/objc/components/video_codec/RTCVideoEncoderH264.mm:
(-[RTCSingleVideoEncoderH264 resetCompressionSessionWithPixelFormat:]):

Source/WebCore:

Covered by existing tests.

* testing/Internals.cpp:
(WebCore::Internals::supportsVCPEncoder):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243811 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoRemove legacy webkitRequestAnimationFrame time quirk
cdumez@apple.com [Wed, 3 Apr 2019 17:27:20 +0000 (17:27 +0000)]
Remove legacy webkitRequestAnimationFrame time quirk
https://bugs.webkit.org/show_bug.cgi?id=196458
<rdar://problem/49490207>

Reviewed by Simon Fraser.

Source/WebCore:

Remove legacy webkitRequestAnimationFrame time quirk and log a deprecation
warning whenever webkitRequestAnimationFrame is called.

* dom/ScriptedAnimationController.cpp:
(WebCore::ScriptedAnimationController::serviceScriptedAnimations):
* page/DOMWindow.cpp:
(WebCore::DOMWindow::requestAnimationFrame):
(WebCore::DOMWindow::webkitRequestAnimationFrame):

LayoutTests:

Rebaseline webkitRequestAnimationFrame layout test now that we log a deprecation
warning.

* fast/animation/request-animation-frame-prefix-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243810 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoResurrect and fix layout test http/tests/adClickAttribution/store-ad-click-attributio...
commit-queue@webkit.org [Wed, 3 Apr 2019 17:23:33 +0000 (17:23 +0000)]
Resurrect and fix layout test http/tests/adClickAttribution/store-ad-click-attribution.html
https://bugs.webkit.org/show_bug.cgi?id=196476

Patch by Alex Christensen <achristensen@webkit.org> on 2019-04-03
Reviewed by Chris Dumez.

Source/WebKit:

Re-apply a change I reverted in r241754 now that it's safe to do so.

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

LayoutTests:

* http/tests/adClickAttribution/store-ad-click-attribution-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243809 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[iOS] Should be able to dismiss picker or popover using the keyboard
commit-queue@webkit.org [Wed, 3 Apr 2019 17:06:04 +0000 (17:06 +0000)]
[iOS] Should be able to dismiss picker or popover using the keyboard
https://bugs.webkit.org/show_bug.cgi?id=196272
<rdar://problem/48943170>

Patch by Daniel Bates <dabates@apple.com> on 2019-04-03
Reviewed by Wenson Hsieh.

Source/WebKit:

Intercept key events and route them to the current input peripheral (if we have one). Add a base key event handler
for all form peripherals that dismisses the accessory when either the Escape key is pressed or Command + . is pressed.
I will fix this issue for the file upload picker/popover in <https://bugs.webkit.org/show_bug.cgi?id=196287>.

* SourcesCocoa.txt: Add file WKFormPeripheralBase.mm.
* UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView endEditingAndUpdateFocusAppearanceWithReason]): Added.
(-[WKContentView resignFirstResponderForWebView]): Write in terms of -endEditingAndUpdateFocusAppearance.
(-[WKContentView inputView]): Code style nit while I am here; add an empty line to demarcate the "crazy"
code that the FIXME is referring to and should ideally be removed from the code that is sane to always do.
(-[WKContentView accessoryDone]): When the accessory is dismissed via the Done button (iPhone) or by pressing
Escape or Command + . using a hardware keyboard (iPhone or iPad) then end the current editing session, but
do not resign first responder status as the page activation state should not be changed.
(-[WKContentView _handleKeyUIEvent:]): Bring back this code when building with USE(UIKIT_KEYBOARD_ADDITIONS)
as we need to route key events to the input peripheral (if we have one). If the input peripheral handles it
then we're done: no need to let UIKit or WebKit handle it when building with USE(UIKIT_KEYBOARD_ADDITIONS),
respectively. If the input peripheral does not handle it then do what we do now.
(-[WKContentView _elementDidFocus:userIsInteracting:blurPreviousNode:changingActivityState:userObject:]):
If the element is re-focused and we have an input peripheral then we want to ensure we are first responder,
reveal the focused element, update the accessory and tell the peripheral that editing has begun (again).
For all other element re-focusing where we don't have a peripheral do what we do now. Also, update _isEditable
to reflect whether the focused element contains selectable text. This is what UIKit wants to know when it queries
-isEditable. Now that we no longer blur the focused element on iPad when the popover is dismissed and keep the
peripheral until there is a focus change we need to ensure that we give the correct answer to UIKit on view
editability. Otherwise, UIKit thinks it needs to update the text selection state when a popup button is tapped
again (as part of its gesture recognizer logic) and this causes an assertion failure in UIKit after it calls back
into us to ask for selection details, which we correctly respond with the equivalent of "we have none" and is
not the answer UIKit expects since we told it we are editable. (Currently we manage to get away with telling UIKit
we are always editable because it is not possible to perform a selection operation when we have a popover open.
Closing the popover blurs the element, setting -isEditable to NO and deallocates the peripheral avoiding this issue).
* UIProcess/ios/forms/WKFormColorControl.h:
* UIProcess/ios/forms/WKFormColorControl.mm:
(-[WKColorPopover controlEndEditing]): Dismiss the popover.
(-[WKFormColorControl initWithView:]): Modified to call base class initializer.
(-[WKFormColorControl assistantView]): Deleted.
(-[WKFormColorControl beginEditing]): Deleted.
(-[WKFormColorControl endEditing]): Deleted.
* UIProcess/ios/forms/WKFormInputControl.h:
* UIProcess/ios/forms/WKFormInputControl.mm:
(-[WKFormInputControl initWithView:]): Modified to call base class initializer.
(-[WKFormInputControl dateTimePickerCalendarType]): Write in terms of self.control.
(-[WKDateTimePopover controlEndEditing]): Dismiss the popover and tell the controller that editing ended.
(-[WKFormInputControl beginEditing]): Deleted.
(-[WKFormInputControl endEditing]): Deleted.
(-[WKFormInputControl assistantView]): Deleted.
* UIProcess/ios/forms/WKFormPeripheral.h:
* UIProcess/ios/forms/WKFormPeripheralBase.h: Added.
* UIProcess/ios/forms/WKFormPeripheralBase.mm: Added.
(-[WKFormPeripheralBase initWithView:control:]): Take ownership of the passed WKFormControl.
(-[WKFormPeripheralBase beginEditing]): Turn around and tell the control.
(-[WKFormPeripheralBase endEditing]): Ditto.
(-[WKFormPeripheralBase assistantView]): Ditto.
(-[WKFormPeripheralBase control]): Return the control.
(-[WKFormPeripheralBase handleKeyEvent:]): Dismiss the accessory (in the same way we dismiss when the Done
button is pressed on iPhone) on keydown of the Escape key or when we receive a UIKeyInputEscape event (for
Command + .).
* UIProcess/ios/forms/WKFormSelectControl.h:
* UIProcess/ios/forms/WKFormSelectControl.mm:
(-[WKFormSelectControl initWithView:]): Modified to call base class initializer.
(-[WKFormSelectControl selectRow:inComponent:extendingSelection:]): Write in terms of self.control.
(-[WKFormSelectControl selectFormPopoverTitle]): Ditto.
(-[WKFormSelectControl assistantView]): Deleted.
(-[WKFormSelectControl beginEditing]): Deleted.
(-[WKFormSelectControl endEditing]): Deleted.
* UIProcess/ios/forms/WKFormSelectPopover.mm:
(-[WKSelectPopover controlEndEditing]): Dismiss the popover.
* WebKit.xcodeproj/project.pbxproj: Add files WKFormPeripheralBase.{h, mm}.

LayoutTests:

Add test to ensure that pressing Escape or Command + . dismisses a picker.

* fast/forms/ios/dismiss-picker-using-keyboard-expected.txt: Added.
* fast/forms/ios/dismiss-picker-using-keyboard.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243808 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoBlob type cannot be stored correctly in IDB when IDBObjectStore has autoIncrement...
sihui_liu@apple.com [Wed, 3 Apr 2019 17:03:59 +0000 (17:03 +0000)]
Blob type cannot be stored correctly in IDB when IDBObjectStore has autoIncrement and keyPath options
https://bugs.webkit.org/show_bug.cgi?id=196128

Reviewed by Geoffrey Garen.

LayoutTests/imported/w3c:

Updated test expectations to PASS.

* web-platform-tests/IndexedDB/nested-cloning-large-expected.txt:
* web-platform-tests/IndexedDB/nested-cloning-large-multiple-expected.txt:
* web-platform-tests/IndexedDB/nested-cloning-small-expected.txt:

Source/WebCore:

If a key is auto-generated, it should become a property of the value object. Network process would perform the
key injection by deserializing IDBValue into script value, setting the property, serializing the result and
storing it in a database record. But network process does not have a JSDOMGlobalObject, so it would fail to
deserialize types including Blob and File.

To solve this issue, we move the key injection to web process and let network process store the original value
it gets. In this case, when web process asks for some value, network process should return key, value and key
path so that web process can decide whether it should perform a key injection before returning the result. Note
that the auto-generated key would always be stored as the key in a ObjectStore record.

Test: storage/indexeddb/modern/objectstore-autoincrement-types.html

* Modules/indexeddb/IDBCursor.cpp:
(WebCore::IDBCursor::setGetResult):
* Modules/indexeddb/IDBCursor.h:
(WebCore::IDBCursor::primaryKeyPath):
* Modules/indexeddb/IDBGetAllResult.cpp:
(WebCore::IDBGetAllResult::isolatedCopy):
(WebCore::IDBGetAllResult::addKey):
(WebCore::IDBGetAllResult::addValue):
(WebCore::IDBGetAllResult::keys const):
(WebCore::IDBGetAllResult::values const):
(WebCore::IDBGetAllResult::allBlobFilePaths const):
(WebCore::isolatedCopyOfVariant): Deleted.

* Modules/indexeddb/IDBGetAllResult.h: Introduce an IDBKeyPath parameter. Also replace Variant with two Vectors,
because we only needed to store either key or value before, and now the stored value could be incomplete.
(WebCore::IDBGetAllResult::IDBGetAllResult):
(WebCore::IDBGetAllResult::keyPath const):
(WebCore::IDBGetAllResult::encode const):
(WebCore::IDBGetAllResult::decode):

* Modules/indexeddb/IDBGetResult.cpp:
(WebCore::IDBGetResult::setValue):
* Modules/indexeddb/IDBGetResult.h:
(WebCore::IDBGetResult::IDBGetResult):
(WebCore::IDBGetResult::keyPath const):
* Modules/indexeddb/IDBObjectStore.cpp:
* Modules/indexeddb/IDBRequest.cpp:
(WebCore::IDBRequest::setResult):
(WebCore::IDBRequest::setResultToStructuredClone):
* Modules/indexeddb/IDBRequest.h:
* Modules/indexeddb/IDBTransaction.cpp:
(WebCore::IDBTransaction::didGetAllRecordsOnServer):
(WebCore::IDBTransaction::didGetRecordOnServer):
* Modules/indexeddb/server/MemoryIDBBackingStore.cpp:
(WebCore::IDBServer::MemoryIDBBackingStore::getRecord):
* Modules/indexeddb/server/MemoryIndex.cpp:
(WebCore::IDBServer::MemoryIndex::getResultForKeyRange const):
(WebCore::IDBServer::MemoryIndex::getAllRecords const):
* Modules/indexeddb/server/MemoryIndexCursor.cpp:
(WebCore::IDBServer::MemoryIndexCursor::currentData):
* Modules/indexeddb/server/MemoryObjectStore.cpp:
(WebCore::IDBServer::MemoryObjectStore::updateIndexesForPutRecord):
(WebCore::IDBServer::MemoryObjectStore::populateIndexWithExistingRecords):
(WebCore::IDBServer::MemoryObjectStore::getAllRecords const):
* Modules/indexeddb/server/MemoryObjectStoreCursor.cpp:
(WebCore::IDBServer::MemoryObjectStoreCursor::currentData):
* Modules/indexeddb/server/SQLiteIDBBackingStore.cpp:
(WebCore::IDBServer::SQLiteIDBBackingStore::updateOneIndexForAddRecord):
(WebCore::IDBServer::SQLiteIDBBackingStore::updateAllIndexesForAddRecord):
(WebCore::IDBServer::SQLiteIDBBackingStore::getRecord):
(WebCore::IDBServer::SQLiteIDBBackingStore::cachedStatementForGetAllObjectStoreRecords):
(WebCore::IDBServer::SQLiteIDBBackingStore::getAllObjectStoreRecords):
(WebCore::IDBServer::SQLiteIDBBackingStore::getAllIndexRecords):
(WebCore::IDBServer::SQLiteIDBBackingStore::getIndexRecord):
(WebCore::IDBServer::SQLiteIDBBackingStore::uncheckedGetIndexRecordForOneKey):
(WebCore::IDBServer::SQLiteIDBBackingStore::openCursor):
(WebCore::IDBServer::SQLiteIDBBackingStore::iterateCursor):
* Modules/indexeddb/server/SQLiteIDBCursor.cpp:
(WebCore::IDBServer::SQLiteIDBCursor::currentData):
* Modules/indexeddb/server/SQLiteIDBCursor.h:

* Modules/indexeddb/server/UniqueIDBDatabase.cpp:
(WebCore::IDBServer::UniqueIDBDatabase::performPutOrAdd): Remove the key injection from network process.
UniqueIDBDatabase stores any value it gets from IDBClient.

* Modules/indexeddb/shared/IDBResultData.cpp:
(WebCore::IDBResultData::getResultRef):
* Modules/indexeddb/shared/IDBResultData.h:

* bindings/js/IDBBindingUtilities.cpp:
(WebCore::injectIDBKeyIntoScriptValue): If property is read-only, set would fail and injectKeyIntoResult would
return null, but we expect it to return result as long as the property value is the same as target. Therefore,
we can add an early return here.
(WebCore::createKeyPathArray):

(WebCore::generateIndexKeyForValue): We used to generate IndexKey from value stored in database but now the
value gets stored does not include auto-generated key, as we remove the key injection from network process. In
this case if the IDBIndex has the same key path as the auto-generated key, IndexKey would be failed to create
for it cannot extract auto-generated key from value. Since the auto-generated key would always be the key in
database record, we could use value of that key when we find a match in key path.

(WebCore::deserializeIDBValueWithKeyInjection): If the key path in the result is single entry, the key is
probably auto-generated, so we could inject the result key into the result value unconditionally.

* bindings/js/IDBBindingUtilities.h:
* bindings/js/JSIDBCursorWithValueCustom.cpp:
(WebCore::JSIDBCursorWithValue::value const):
* bindings/js/JSIDBRequestCustom.cpp:
(WebCore::JSIDBRequest::result const):

LayoutTests:

* TestExpectations:
* storage/indexeddb/modern/objectstore-autoincrement-types-expected.txt: Added.
* storage/indexeddb/modern/objectstore-autoincrement-types.html: Added.
* storage/indexeddb/modern/resources/objectstore-autoincrement-types.js: Added.
(next):
(prepareDatabase.event.target.onsuccess):
(prepareDatabase):
(testSteps):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243807 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoResetting quota should take into account third party origins
youenn@apple.com [Wed, 3 Apr 2019 17:00:36 +0000 (17:00 +0000)]
Resetting quota should take into account third party origins
https://bugs.webkit.org/show_bug.cgi?id=196462

Reviewed by Geoffrey Garen.

Source/WebKit:

When clearing the storage quota, we were resetting it to the default value
without taking care of whether third party or not.
Updated the code to ensure that the default quota is computed based on the origin.

Updated existing test to cover this case.

* NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::clearStorageQuota):
(WebKit::NetworkProcess::setCacheStorageParameters):
(WebKit::NetworkProcess::updateQuotaBasedOnSpaceUsageForTesting):
(WebKit::NetworkProcess::storageQuotaManager):
* NetworkProcess/NetworkProcess.h:
(WebKit::NetworkProcess::StorageQuotaManagers::defaultQuota const):
(WebKit::NetworkProcess::StorageQuotaManagers::setDefaultQuotas):
(WebKit::NetworkProcess::StorageQuotaManagers::managersPerOrigin):

LayoutTests:

* http/wpt/cache-storage/quota-third-party.https-expected.txt:
* http/wpt/cache-storage/quota-third-party.https.html:
* platform/mac-wk2/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243806 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[GTK] Unreviewed gardening, update test expectations
dpino@igalia.com [Wed, 3 Apr 2019 16:53:21 +0000 (16:53 +0000)]
[GTK] Unreviewed gardening, update test expectations

* platform/gtk/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243805 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoGet rid of HTMLInputElement::setEditingValue
mcatanzaro@igalia.com [Wed, 3 Apr 2019 16:40:16 +0000 (16:40 +0000)]
Get rid of HTMLInputElement::setEditingValue
https://bugs.webkit.org/show_bug.cgi?id=196402

Reviewed by Darin Adler.

Source/WebCore:

HTMLInputElement::setEditingValue is only used for Epiphany password autofill. We did it
this way because that's what Chrome uses for autofill, but Apple uses
HTMLInputElement::setValueForUser. Let's switch to that instead, then we can get rid of
setEditingValue.

This fixes logging into ting.com after username and password are autofilled by Epiphany.
Before this change, the login would fail unless you first manually edit either the username
or the password field.

* html/HTMLInputElement.cpp:
(WebCore::HTMLInputElement::setEditingValue): Deleted.
* html/HTMLInputElement.h:
* testing/Internals.cpp:
(WebCore::Internals::setEditingValue): Deleted.
* testing/Internals.h:
* testing/Internals.idl:

Source/WebKit:

* WebProcess/InjectedBundle/API/glib/DOM/WebKitDOMElement.cpp:
(webkit_dom_element_html_input_element_set_editing_value):
* WebProcess/InjectedBundle/API/gtk/DOM/WebKitDOMHTMLInputElement.cpp:
(webkit_dom_html_input_element_set_editing_value):

LayoutTests:

* fast/forms/editing-value-expected.txt: Removed.
* fast/forms/editing-value-null-renderer-expected.txt: Removed.
* fast/forms/editing-value-null-renderer.html: Removed.
* fast/forms/editing-value.html: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243804 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[ews] Clean out ~/Library/Logs/CoreSimulator/ on reboot
jbedard@apple.com [Wed, 3 Apr 2019 16:39:58 +0000 (16:39 +0000)]
[ews] Clean out ~/Library/Logs/CoreSimulator/ on reboot
https://bugs.webkit.org/show_bug.cgi?id=196540

Reviewed by Aakash Jain.

* EWSTools/start-queue-mac.sh: Remove CoreSimulator logs.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243803 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoThe following layout tests are flaky failures
sroberts@apple.com [Wed, 3 Apr 2019 16:30:15 +0000 (16:30 +0000)]
The following layout tests are flaky failures
imported/w3c/web-platform-tests/html/semantics/embedded-content/the-img-element/sizes/parse-a-sizes-attribute-width-1000px.html
imported/w3c/web-platform-tests/html/semantics/embedded-content/media-elements/ready-states/autoplay.html
imported/w3c/web-platform-tests/html/semantics/embedded-content/media-elements/error-codes/error.html
https://bugs.webkit.org/show_bug.cgi?id=195466

Unreviewed test gardening.

* platform/ios-simulator/TestExpectations: Updating test expectations for flaky failures

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243802 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoLayout Test http/wpt/resource-timing/rt-resources-per-worker.html is flaky
sroberts@apple.com [Wed, 3 Apr 2019 16:14:07 +0000 (16:14 +0000)]
Layout Test http/wpt/resource-timing/rt-resources-per-worker.html is flaky
https://bugs.webkit.org/show_bug.cgi?id=180260

Unreviewed test gardening.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243801 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoAdd a WebKit internal feature flag to always enable modern compatibility mode by...
wenson_hsieh@apple.com [Wed, 3 Apr 2019 16:02:06 +0000 (16:02 +0000)]
Add a WebKit internal feature flag to always enable modern compatibility mode by default
https://bugs.webkit.org/show_bug.cgi?id=196526
<rdar://problem/49532923>

Reviewed by Antoine Quint.

* Shared/WebPreferences.yaml:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243800 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoMissing includes that were previously provided via UnifiedSources
commit-queue@webkit.org [Wed, 3 Apr 2019 15:58:04 +0000 (15:58 +0000)]
Missing includes that were previously provided via UnifiedSources
https://bugs.webkit.org/show_bug.cgi?id=196434

Unreviewed build fix.

Patch by Pablo Saavedra <psaavedra@igalia.com> on 2019-04-03

* html/InputType.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243799 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoIntroduce and add plumbing for a website policy for meta viewport tag handling
wenson_hsieh@apple.com [Wed, 3 Apr 2019 15:22:10 +0000 (15:22 +0000)]
Introduce and add plumbing for a website policy for meta viewport tag handling
https://bugs.webkit.org/show_bug.cgi?id=196285

Reviewed by Tim Horton.

Source/WebCore:

Add MetaViewportPolicy to DocumentLoader. See WebKit ChangeLog for more detail.

* loader/DocumentLoader.h:
(WebCore::DocumentLoader::metaViewportPolicy const):
(WebCore::DocumentLoader::setMetaViewportPolicy):

Source/WebKit:

Add WebsiteMetaViewportPolicy, a bit that can be used to determine whether to respect or ignore the meta
viewport tag and use native web page parameters instead of the default parameters.

* Shared/WebsiteMetaViewportPolicy.h: Copied from Source/WebKit/Shared/WebsitePoliciesData.h.
* Shared/WebsitePoliciesData.cpp:
(WebKit::WebsitePoliciesData::encode const):
(WebKit::WebsitePoliciesData::decode):
(WebKit::WebsitePoliciesData::applyToDocumentLoader):

Convert WebKit::WebsiteMetaViewportPolicy into WebCore::MetaViewportPolicy.

* Shared/WebsitePoliciesData.h:
* UIProcess/API/APIWebsitePolicies.cpp:
(API::WebsitePolicies::data):
* UIProcess/API/APIWebsitePolicies.h:

Add additional plumbing for the policy flag.

* WebKit.xcodeproj/project.pbxproj:
* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::resetViewportDefaultConfiguration):

Use native web page parameters if either "shouldIgnoreMetaViewport" is enabled, or the new policy is set.
Eventually, the policy should completely replace the former preference once no internal clients depend on it.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243798 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoAdd plumbing for a compatibility mode preference in WebKit
wenson_hsieh@apple.com [Wed, 3 Apr 2019 15:02:07 +0000 (15:02 +0000)]
Add plumbing for a compatibility mode preference in WebKit
https://bugs.webkit.org/show_bug.cgi?id=196005

Reviewed by Tim Horton.

Source/WebKit:

Add plumbing between the Cocoa API object (WKWebpagePreferences) and the inner C++ API object
(API::WebsitePolicies) for compatibility mode.

* Shared/WebCompatibilityMode.h: Added.
* UIProcess/API/APIWebsitePolicies.h:
* UIProcess/API/Cocoa/WKWebpagePreferences.mm:
* UIProcess/Cocoa/NavigationState.mm:
(WebKit::NavigationState::NavigationClient::decidePolicyForNavigationAction):
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::adjustPoliciesForCompatibilityMode):
* UIProcess/WebPageProxy.h:
* UIProcess/ios/WebPageProxyIOS.mm:
* WebKit.xcodeproj/project.pbxproj:

Tools:

Add new API tests.

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/ios/PreferredCompatibilityMode.mm: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243797 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[CoordinatedGraphics] Hidden pages are not suspended after a web view resize
carlosgc@webkit.org [Wed, 3 Apr 2019 13:00:16 +0000 (13:00 +0000)]
[CoordinatedGraphics] Hidden pages are not suspended after a web view resize
https://bugs.webkit.org/show_bug.cgi?id=196487

Reviewed by Žan Doberšek.

When resizing the window, the hidden tabs are updated too, to avoid flickering or getting the old size when
switching tabs. For that we need to resume painting but we are not suspending it again after the update.

* WebProcess/WebPage/CoordinatedGraphics/DrawingAreaCoordinatedGraphics.cpp:
(WebKit::DrawingAreaCoordinatedGraphics::updateBackingStoreState): Suspend the painting again after a
synchronous update if needed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243796 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoWeb Inspector: Remote Inspector indicate callback should always happen on the main...
pecoraro@apple.com [Wed, 3 Apr 2019 12:42:38 +0000 (12:42 +0000)]
Web Inspector: Remote Inspector indicate callback should always happen on the main thread
https://bugs.webkit.org/show_bug.cgi?id=196513
<rdar://problem/49498284>

Reviewed by Devin Rousso.

Source/JavaScriptCore:

* inspector/remote/cocoa/RemoteInspectorCocoa.mm:
(Inspector::RemoteInspector::receivedIndicateMessage):
When we have a WebThread, don't just run on the WebThread,
run on the MainThread with the WebThreadLock.

Source/WebCore:

* platform/ios/wak/WebCoreThreadSystemInterface.cpp:
(InitWebCoreThreadSystemInterface):

Source/WTF:

* wtf/MainThread.h:
* wtf/cocoa/MainThreadCocoa.mm:
(WTF::dispatchAsyncOnMainThreadWithWebThreadLockIfNeeded):
* wtf/ios/WebCoreThread.cpp:
* wtf/ios/WebCoreThread.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243795 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[WPE][GTK] Update libsrtp dependency package for Debian
csaavedra@igalia.com [Wed, 3 Apr 2019 08:33:41 +0000 (08:33 +0000)]
[WPE][GTK] Update libsrtp dependency package for Debian
https://bugs.webkit.org/show_bug.cgi?id=196528

Reviewed by Frédéric Wang.

* gtk/install-dependencies:
* wpe/install-dependencies:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243794 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoWeb Inspector: [GTK] Copy copies to nowhere
carlosgc@webkit.org [Wed, 3 Apr 2019 08:25:26 +0000 (08:25 +0000)]
Web Inspector: [GTK] Copy copies to nowhere
https://bugs.webkit.org/show_bug.cgi?id=181228

Reviewed by Michael Catanzaro.

The thing is that those items are actually submenu items, with options, for example in the case of Copy to copy
HTML, Text, XPath, etc. We are not correctly handling submenus when populating the context menu received from
the web process.

* UIProcess/gtk/WebContextMenuProxyGtk.cpp:
(WebKit::WebContextMenuProxyGtk::populateSubMenu): Helper to populate submenu items recursively.
(WebKit::WebContextMenuProxyGtk::populate): Handle submenu items.
* UIProcess/gtk/WebContextMenuProxyGtk.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243793 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[Cocoa] Add new API around WKWebpagePreferences in WKNavigationDelegate and WKWebView...
wenson_hsieh@apple.com [Wed, 3 Apr 2019 04:58:04 +0000 (04:58 +0000)]
[Cocoa] Add new API around WKWebpagePreferences in WKNavigationDelegate and WKWebViewConfiguration
https://bugs.webkit.org/show_bug.cgi?id=196284
<rdar://problem/47228232>

Reviewed by Tim Horton.

Tests for both of these APIs will be added in a subsequent patch.

* Shared/API/Cocoa/WebKit.h:
* UIProcess/API/APIPageConfiguration.cpp:
(API::PageConfiguration::defaultWebsitePolicies const):
(API::PageConfiguration::setDefaultWebsitePolicies):
* UIProcess/API/APIPageConfiguration.h:
* UIProcess/API/Cocoa/WKNavigationDelegate.h:

Add a new navigation delegate hook to allow clients to return a WKWebpagePreference targeting the given
navigation action.

* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _initializeWithConfiguration:]):
* UIProcess/API/Cocoa/WKWebViewConfiguration.h:
* UIProcess/API/Cocoa/WKWebViewConfiguration.mm:

Add new API on WKWebViewConfiguration to specify a default WKWebpagePreference to use when navigating.

(-[WKWebViewConfiguration copyWithZone:]):
(-[WKWebViewConfiguration defaultWebpagePreferences]):
(-[WKWebViewConfiguration setDefaultWebpagePreferences:]):
* UIProcess/API/Cocoa/WKWebpagePreferences.h:
* UIProcess/Cocoa/NavigationState.h:
* UIProcess/Cocoa/NavigationState.mm:
(WebKit::NavigationState::setNavigationDelegate):
(WebKit::NavigationState::NavigationClient::decidePolicyForNavigationAction):

Invoke the new WKWebpagePreferences-based navigation delegate method.

* WebKit.xcodeproj/project.pbxproj:
* mac/postprocess-framework-headers.sh:
* mac/replace-webkit-additions-in-framework-headers.sh: Added.

Move logic in the "Postprocess Framework Headers" step responsible for stripping away included files from
WebKitAdditions out into a separate build phase, called "Replace WebKitAdditions in Framework Headers". This
ensures headers attempting to include from WebKitAdditions have these additional statements removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243787 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoREGRESSION (r238266): Exchange 2013 Outlook Web Access displays partially blank page...
simon.fraser@apple.com [Wed, 3 Apr 2019 04:20:24 +0000 (04:20 +0000)]
REGRESSION (r238266): Exchange 2013 Outlook Web Access displays partially blank page when creating new e-mail
https://bugs.webkit.org/show_bug.cgi?id=196522
Source/WebCore:

rdar://problem/49472941

Reviewed by Zalan Bujtas.

In this content a layer is composited to clip descendants, and has negative z-order children,
so we compute that it "paints into ancestor", and has a foreground layer. This combination doesn't
make sense, and when the layer becomes scrollable, we end up with bad paint phases on layers, and
fail to paint the contents.

Fix by ensuring that a layer has its own backing store if it requires a foreground layer
by virtue of having negative z-order children.

Test: compositing/backing/foreground-layer-no-paints-into-ancestor.html

* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::requiresOwnBackingStore const):

LayoutTests:

Reviewed by Zalan Bujtas.

* compositing/backing/foreground-layer-no-paints-into-ancestor-expected.html: Added.
* compositing/backing/foreground-layer-no-paints-into-ancestor.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243786 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoCursor count badge reverts to Zero during Drag & Drop of multiple items.
timothy@apple.com [Wed, 3 Apr 2019 03:46:37 +0000 (03:46 +0000)]
Cursor count badge reverts to Zero during Drag & Drop of multiple items.
https://bugs.webkit.org/show_bug.cgi?id=196511

Reviewed by Daniel Bates.

* page/DragController.cpp:
(WebCore::DragController::tryDocumentDrag): Don't set m_numberOfItemsToBeAccepted to
zero when dragging to a non-file input element.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243785 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoEliminate plugin sandbox exceptions
ggaren@apple.com [Wed, 3 Apr 2019 03:24:29 +0000 (03:24 +0000)]
Eliminate plugin sandbox exceptions
https://bugs.webkit.org/show_bug.cgi?id=196510

Reviewed by Chris Dumez.

* PluginProcess/mac/PluginProcessMac.mm:
(WebKit::PluginProcess::initializeSandbox):
* UIProcess/Plugins/PluginInfoStore.cpp:
(WebKit::PluginInfoStore::shouldAllowPluginToRunUnsandboxed): Deleted.
* UIProcess/Plugins/PluginInfoStore.h:
* UIProcess/Plugins/mac/PluginInfoStoreMac.mm:
(WebKit::PluginInfoStore::shouldUsePlugin):
(WebKit::PluginInfoStore::shouldAllowPluginToRunUnsandboxed): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243784 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[WK2] Add support for Window's beforeprint / afterprint events
Hironori.Fujii@sony.com [Wed, 3 Apr 2019 01:56:58 +0000 (01:56 +0000)]
[WK2] Add support for Window's beforeprint / afterprint events
https://bugs.webkit.org/show_bug.cgi?id=196478
<rdar://problem/49535124>

Unreviewed build fix for MSVC.

> WebKitTestRunner\TestController.cpp(267): error C2397: conversion from 'double' to 'float' requires a narrowing conversion

And, there is a style guideline. <https://webkit.org/code-style-guidelines/#floating-point-literals>

* WebKitTestRunner/TestController.cpp:
(WTR::printFrame): Use int and float literals.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243783 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoHTML Parser: Remove conditional parsing of <noembed> content
cdumez@apple.com [Wed, 3 Apr 2019 01:55:38 +0000 (01:55 +0000)]
HTML Parser: Remove conditional parsing of <noembed> content
https://bugs.webkit.org/show_bug.cgi?id=196514

Reviewed by Geoffrey Garen.

LayoutTests/imported/w3c:

Resync WPT after https://github.com/web-platform-tests/wpt/pull/15471 to gain
test coverage. Both Gecko and Blink are passing the new check, only WebKit was
failing.

* web-platform-tests/domparsing/DOMParser-parseFromString-html-expected.txt:
* web-platform-tests/domparsing/DOMParser-parseFromString-html.html:

Source/WebCore:

Our HTML Parser has raw text handling for <noembed> content only if plugins are runnable.
However, the HTML specification doesn't ask such behavior [1], and it doesn't match to
our HTML serializer. We should always handle it as raw text.

Blink already made this change in https://chromium-review.googlesource.com/c/1477556.

[1] https://html.spec.whatwg.org/multipage/parsing.html#parsing-html-fragments:noembed

No new tests, updated existing test.

* html/parser/HTMLParserOptions.cpp:
(WebCore::HTMLParserOptions::HTMLParserOptions):
* html/parser/HTMLParserOptions.h:
* html/parser/HTMLTokenizer.cpp:
(WebCore::HTMLTokenizer::updateStateFor):
* html/parser/HTMLTreeBuilder.cpp:
(WebCore::HTMLTreeBuilder::processStartTagForInBody):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243782 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoWeb Inspector: Elements: DOM breakpoint icon should stroke when hovering selection...
drousso@apple.com [Wed, 3 Apr 2019 00:44:42 +0000 (00:44 +0000)]
Web Inspector: Elements: DOM breakpoint icon should stroke when hovering selection area
https://bugs.webkit.org/show_bug.cgi?id=196516
<rdar://problem/49540494>

Reviewed by Timothy Hatcher.

* UserInterface/Views/DOMTreeContentView.css:
(body:not(.window-inactive) .content-view.dom-tree .tree-outline.dom:focus li:matches(.selected, .hovered) .status-image.breakpoint,): Added.
(body:not(.window-inactive) .content-view.dom-tree .tree-outline.dom:focus li:matches(.selected, .hovered) .status-image.breakpoint.subtree): Added.
(body:not(.window-inactive) .content-view.dom-tree .tree-outline.dom:focus li:matches(.selected, :hover) .status-image.breakpoint): Deleted.
(body:not(.window-inactive) .content-view.dom-tree .tree-outline.dom:focus li:matches(.selected, :hover) .status-image.breakpoint.subtree): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243781 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoSVG Animation (SMIL) on <text> or <tspan> doesn't work on second run
commit-queue@webkit.org [Wed, 3 Apr 2019 00:26:04 +0000 (00:26 +0000)]
SVG Animation (SMIL) on <text> or <tspan> doesn't work on second run
https://bugs.webkit.org/show_bug.cgi?id=150388

Patch by Said Abou-Hallawa <sabouhallawa@apple.com> on 2019-04-02
Reviewed by Simon Fraser.

The bug is fixed by removing the SVG tear off objects. See
https://bugs.webkit.org/show_bug.cgi?id=191237.

We need to add a layout test to ensure that no code change will break
this scenario in the future.

* svg/animations/smil-multiple-animate-list-expected.svg: Added.
* svg/animations/smil-multiple-animate-list.svg: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243780 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoimported/w3c/web-platform-tests/mst-content-hint/MediaStreamTrack-contentHint.html...
sroberts@apple.com [Wed, 3 Apr 2019 00:04:48 +0000 (00:04 +0000)]
imported/w3c/web-platform-tests/mst-content-hint/MediaStreamTrack-contentHint.html is a flaky crash
https://bugs.webkit.org/show_bug.cgi?id=196517

Unreviewed test gardening.

* platform/mac/TestExpectations: Skipping test for flaky crashing

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243779 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[Mojave] Layout test fast/text/complex-initial-advance.html is failing
ryanhaddad@apple.com [Tue, 2 Apr 2019 23:33:00 +0000 (23:33 +0000)]
[Mojave] Layout test fast/text/complex-initial-advance.html is failing
https://bugs.webkit.org/show_bug.cgi?id=196512

Unreviewed test gardening.

* platform/mac/TestExpectations: Mark test as failing.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243768 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoFix assertion in http/tests/adClickAttribution/store-ad-click-attribution.html
achristensen@apple.com [Tue, 2 Apr 2019 23:11:43 +0000 (23:11 +0000)]
Fix assertion in http/tests/adClickAttribution/store-ad-click-attribution.html
https://bugs.webkit.org/show_bug.cgi?id=196503

Reviewed by Chris Dumez.

Source/WebKit:

In WebPageProxy::didDestroyNavigation we try to ignore a request to destroy a navigation from a page being navigated from
during a cross-site navigation, but if the old web process sends the message after WebPageProxy::commitProvisionalPage
has been called, we can still destroy a navigation when we are continuing a navigation in another process.  To prevent this,
have the process not send the message when it knows the navigation is continuing in another process.
Also make the use of unchecked navigation pointers more robust by checking it for nullity.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::didCommitLoadForFrame):
* WebProcess/WebPage/WebFrame.cpp:
(WebKit::WebFrame::didReceivePolicyDecision):
(WebKit::WebFrame::documentLoaderDetached):
* WebProcess/WebPage/WebFrame.h:

LayoutTests:

* http/tests/adClickAttribution/store-ad-click-attribution-expected.txt:
* http/tests/adClickAttribution/store-ad-click-attribution.html:
* platform/wk2/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243767 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[ews-build] Reassign bots to queues
aakash_jain@apple.com [Tue, 2 Apr 2019 23:07:49 +0000 (23:07 +0000)]
[ews-build] Reassign bots to queues

* BuildSlaveSupport/ews-build/config.json: Reassigned ews150 and ews152.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243766 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoXMLHttpRequestUpload's loadstart event not correct initialized
cdumez@apple.com [Tue, 2 Apr 2019 23:00:47 +0000 (23:00 +0000)]
XMLHttpRequestUpload's loadstart event not correct initialized
https://bugs.webkit.org/show_bug.cgi?id=196174
<rdar://problem/49191412>

Reviewed by Alex Christensen.

LayoutTests/imported/w3c:

* web-platform-tests/xhr/event-error-order.sub.html:
Update test after https://github.com/web-platform-tests/wpt/pull/13365

* web-platform-tests/xhr/abort-during-upload-expected.txt:
* web-platform-tests/xhr/event-error-order.sub-expected.txt:
* web-platform-tests/xhr/event-loadstart-upload-expected.txt:
* web-platform-tests/xhr/event-timeout-order-expected.txt:
* web-platform-tests/xhr/send-response-event-order-expected.txt:
Rebaseline several WPT tests that are now passing.

Source/WebCore:

Align progress event firing with the XHR specification.

No new tests, rebaselined existing tests.

* xml/XMLHttpRequest.cpp:
(WebCore::XMLHttpRequest::createRequest):
As per [1], the loadstart event fired on the XMLHttpRequestUpload object should use
loaded=0 and total=`req’s body’s total bytes`.
[1] https://xhr.spec.whatwg.org/#the-send()-method (step 11.2.)

(WebCore::XMLHttpRequest::didSendData):
As per [2], the progress / load / loadend should use loaded=transmitted and total=length.
[2] https://xhr.spec.whatwg.org/#ref-for-process-request-end-of-body (steps 5, 6 and 7)

(WebCore::XMLHttpRequest::didReceiveData):
As per [3], we should fire the readystatechange event *before* the progress event.
This is covered by web-platform-tests/xhr/send-response-event-order.htm which was failing
differently after the other changes in this patch.
[3] https://xhr.spec.whatwg.org/#ref-for-process-response (steps 9.4 and 9.5)

(WebCore::XMLHttpRequest::dispatchErrorEvents):
As per [4], in case of an error, we should fire the provided 'event' and 'loadend' with
loaded=0 and total=0.
[4] https://xhr.spec.whatwg.org/#request-error-steps (steps 7 and 8)

* xml/XMLHttpRequestUpload.cpp:
(WebCore::XMLHttpRequestUpload::dispatchProgressEvent):
* xml/XMLHttpRequestUpload.h:
Simplify XMLHttpRequestUpload. It no longer needs to store loaded / total as data
members now that they are always passed by the call site. lengthComputable is set
to !!total as [5] says to set it to true if length/total is not 0.
[5] https://xhr.spec.whatwg.org/#concept-event-fire-progress

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243765 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[iOS] Fix iokit-get-properties sandbox violation in the WebContent process
pvollan@apple.com [Tue, 2 Apr 2019 22:39:11 +0000 (22:39 +0000)]
[iOS] Fix iokit-get-properties sandbox violation in the WebContent process
https://bugs.webkit.org/show_bug.cgi?id=196500
<rdar://problem/49497788>

Reviewed by Brent Fulgham.

Attempting to get the iokit properties 'product-description', 'artwork-display-gamut' and
'artwork-scale-factor' is causing sandbox violations.

* Resources/SandboxProfiles/ios/com.apple.WebKit.WebContent.sb:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243764 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoWeb Inspector: Canvas: add support for showing WebGPU contexts
drousso@apple.com [Tue, 2 Apr 2019 22:19:05 +0000 (22:19 +0000)]
Web Inspector: Canvas: add support for showing WebGPU contexts
https://bugs.webkit.org/show_bug.cgi?id=196413
<rdar://problem/49438898>

Reviewed by Timothy Hatcher.

Source/WebCore:

Tests: inspector/canvas/create-context-webgpu.html
       inspector/canvas/resolveCanvasContext-webgpu.html

* Modules/webgpu/GPUCanvasContext.idl:
* Modules/webgpu/GPUCanvasContext.h:
(WebCore::GPUCanvasContext::canvas const): Added.
* Modules/webgpu/GPUCanvasContext.cpp:
(WebCore::GPUCanvasContext::create):

* testing/InternalSettings.idl:
* testing/InternalSettings.h:
* testing/InternalSettings.cpp:
(WebCore::InternalSettings::setWebGPUEnabled): Added.

LayoutTests:

* inspector/canvas/create-context-webgpu.html: Added.
* inspector/canvas/create-context-webgpu-expected.txt: Added.
* inspector/canvas/resolveCanvasContext-webgpu.html: Added.
* inspector/canvas/resolveCanvasContext-webgpu-expected.txt: Added.

* inspector/canvas/resources/create-context-utilities.js:
(createDetachedCanvas):
(createCSSCanvas):
(destroyCanvases):
(TestPage.registerInitializer.InspectorTest.CreateContextUtilities.initializeTestSuite): Added.
(TestPage.registerInitializer.InspectorTest.CreateContextUtilities.addSimpleTestCase): Added.
(TestPage.registerInitializer.InspectorTest.CreateContextUtilities.addCSSCanvasTestCase): Added.
(TestPage.registerInitializer.window.initializeTestSuite): Deleted.
(TestPage.registerInitializer.window.addSimpleTestCase): Deleted.
(TestPage.registerInitializer.window.addCSSCanvasTestCase): Deleted.

* inspector/canvas/create-context-2d.html:
* inspector/canvas/create-context-2d-expected.txt:
* inspector/canvas/create-context-bitmaprenderer.html:
* inspector/canvas/create-context-bitmaprenderer-expected.txt:
* inspector/canvas/create-context-webgl.html:
* inspector/canvas/create-context-webgl-expected.txt:
* inspector/canvas/create-context-webgl2.html:
* inspector/canvas/create-context-webgl2-expected.txt:

* platform/gtk/TestExpectations:
* platform/ios/TestExpectations:
* platform/mac-wk1/TestExpectations:
* platform/mac/TestExpectations:
* platform/win/TestExpectations:
* platform/wincairo/TestExpectations:
* platform/wpe/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243763 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[WK2] Add support for Window's beforeprint / afterprint events
cdumez@apple.com [Tue, 2 Apr 2019 21:43:24 +0000 (21:43 +0000)]
[WK2] Add support for Window's beforeprint / afterprint events
https://bugs.webkit.org/show_bug.cgi?id=196478

Reviewed by Alex Christensen.

LayoutTests/imported/w3c:

Rebaseline WPT tests now that more checks are passing.

* web-platform-tests/html/browsers/the-window-object/security-window/window-security.https-expected.txt:
* web-platform-tests/html/browsers/the-window-object/security-window/window-security.sub-expected.txt:
* web-platform-tests/html/browsers/the-window-object/window-properties.https-expected.txt:
* web-platform-tests/html/dom/interfaces-expected.txt:
* web-platform-tests/html/webappapis/scripting/events/body-exposed-window-event-handlers-expected.txt:
* web-platform-tests/html/webappapis/scripting/events/event-handler-attributes-body-window-expected.txt:

Source/WebCore:

Add support for Window's beforeprint / afterprint events as per:
- https://html.spec.whatwg.org/#dom-print

Blink and Gecko already support this.

Test: printing/printing-events.html

* dom/EventNames.h:
* html/HTMLAttributeNames.in:
* html/HTMLBodyElement.cpp:
(WebCore::HTMLBodyElement::createWindowEventHandlerNameMap):
* page/Page.cpp:
(WebCore::dispatchPrintEvent):
(WebCore::Page::dispatchBeforePrintEvent):
(WebCore::Page::dispatchAfterPrintEvent):
* page/Page.h:
* page/WindowEventHandlers.idl:

Source/WebKit:

Add support for Window's beforeprint / afterprint events as per:
- https://html.spec.whatwg.org/#dom-print

Blink and Gecko already support this.

* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::beginPrinting):
(WebKit::WebPage::endPrinting):

Tools:

Add minimal implementation for printFrame in WebKitTestRunner.

* WebKitTestRunner/TestController.cpp:
(WTR::printFrame):
(WTR::TestController::createOtherPage):
(WTR::TestController::createWebViewWithOptions):

LayoutTests:

Add layout test coverage.

* TestExpectations:
* platform/wk2/TestExpectations:
* printing/printing-events-expected.txt: Added.
* printing/printing-events.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243762 268f45cc-cd09-0410-ab3c-d52691b4dbfc