WebKit-https.git
15 months agoAdd option to run-api-tests to force running of DISABLED tests.
jer.noble@apple.com [Thu, 16 Aug 2018 22:37:22 +0000 (22:37 +0000)]
Add option to run-api-tests to force running of DISABLED tests.
https://bugs.webkit.org/show_bug.cgi?id=188674

Reviewed by Joseph Pecoraro.

* Scripts/webkitpy/api_tests/run_api_tests.py:
(parse_args):
* Scripts/webkitpy/api_tests/runner.py:
(Runner.command_for_port):
(_Worker._run_single_test):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234962 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[ews-build] Separate queues for Builders and Testers - macOS High Sierra
aakash_jain@apple.com [Thu, 16 Aug 2018 22:11:48 +0000 (22:11 +0000)]
[ews-build] Separate queues for Builders and Testers - macOS High Sierra
https://bugs.webkit.org/show_bug.cgi?id=188540

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-build/config.json: Splitted few macOS queues into builders and testers.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234961 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoAdd temporary SPI WKContextHandlesSafeBrowsing
achristensen@apple.com [Thu, 16 Aug 2018 22:10:20 +0000 (22:10 +0000)]
Add temporary SPI WKContextHandlesSafeBrowsing
https://bugs.webkit.org/show_bug.cgi?id=188676

Reviewed by Joseph Pecoraro.

WebKit showing the safe browsing warning doesn't play well with Safari showing the safe browsing warning.
I plan to adopt this SPI in Safari to disable Safari's safe browsing check if it's true.
Then when I implement safe browsing in WebKit, I can switch this value to true in the same change and
not have a broken Safari.

* UIProcess/API/C/mac/WKContextPrivateMac.h:
* UIProcess/API/C/mac/WKContextPrivateMac.mm:
(WKContextHandlesSafeBrowsing):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234960 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoAdd script to generate WebContent service resource files
commit-queue@webkit.org [Thu, 16 Aug 2018 21:59:05 +0000 (21:59 +0000)]
Add script to generate WebContent service resource files
https://bugs.webkit.org/show_bug.cgi?id=188601

Patch by Ben Richards <benton_richards@apple.com> on 2018-08-16
Reviewed by Dan Bernstein.

Added new build phase to WebContent service to copy resource files to WebKit.framework/PrivateHeaders/CustomWebContentResource.
These resources are intended to be used by a client to create a custom WebContent service.

* Scripts/copy-webcontent-resources-to-private-headers.sh: Added.
* WebKit.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234958 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoCustom element constructor doesn't use HTMLElement in new.target's realm
rniwa@webkit.org [Thu, 16 Aug 2018 21:18:02 +0000 (21:18 +0000)]
Custom element constructor doesn't use HTMLElement in new.target's realm
https://bugs.webkit.org/show_bug.cgi?id=188634

Reviewed by Keith Miller.

LayoutTests/imported/w3c:

Rebaselined the test now that all relevant test cases pass. All remaining test failures are
for customized builtin, which we do not and shall not implement.

* web-platform-tests/custom-elements/htmlconstructor/newtarget-expected.txt:

Source/WebCore:

Fixed the bug that HTMLElement's constructor was constructing an element of its own realm
instead of the realm of new.target. This results in the JS wrapper created for the element
belonging to the global object of the HTMLElement constructor which was invoked instead of
the global object of new.target as specified in:
https://html.spec.whatwg.org/multipage/dom.html#html-element-constructors

In particular, step 9.2. specifies that we "perform element.[[SetPrototypeOf]](prototype)."
where prototype is the result of Get(NewTarget, "prototype") in step 7.

WebKit's new behavior matches that of Chrome and Firefox.

Test: imported/w3c/web-platform-tests/custom-elements/htmlconstructor/newtarget.html

* bindings/js/JSHTMLElementCustom.cpp:
(WebCore::constructJSHTMLElement):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234957 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoUnreviewed, fix the watchOS build after r188464.
akeerthi@apple.com [Thu, 16 Aug 2018 21:10:57 +0000 (21:10 +0000)]
Unreviewed, fix the watchOS build after r188464.

* platform/ios/PlatformPasteboardIOS.mm:
(WebCore::PlatformPasteboard::setColor):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234956 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[Curl] Test gardening
Basuke.Suzuki@sony.com [Thu, 16 Aug 2018 21:09:26 +0000 (21:09 +0000)]
[Curl] Test gardening
https://bugs.webkit.org/show_bug.cgi?id=188675

Unreviewed test gardening.

Fix expectations to match after r234892 landed.

* platform/wincairo/TestExpectations:
* platform/wincairo/http/tests/xmlhttprequest/methods-async-expected.txt: Added
* platform/wincairo/http/tests/xmlhttprequest/methods-expected.txt:
* platform/wincairo/http/tests/xmlhttprequest/workers/methods-async-expected.txt: Added
* platform/wincairo/http/tests/xmlhttprequest/workers/methods-expected.txt: Added

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234955 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoConsolidate data/string API loading paths
achristensen@apple.com [Thu, 16 Aug 2018 20:52:30 +0000 (20:52 +0000)]
Consolidate data/string API loading paths
https://bugs.webkit.org/show_bug.cgi?id=188417

Reviewed by Michael Catanzaro.

loadHTMLString and loadData are basically duplicate code.
loadPlainTextString was also basically the same except it didn't set up a navigation, which
was almost certainly a bug, but nobody uses it in all of Apple and Debian.  We should probably deprecate
and remove it, but for now I make it use the same data loading path.

* UIProcess/API/C/WKPage.cpp:
(WKPageLoadData):
(WKPageLoadDataWithUserData):
(loadString):
(WKPageLoadHTMLString):
(WKPageLoadHTMLStringWithUserData):
(WKPageLoadPlainTextString):
(WKPageLoadPlainTextStringWithUserData):
* UIProcess/API/Cocoa/WKBrowsingContextController.mm:
(-[WKBrowsingContextController loadHTMLString:baseURL:userData:]):
(-[WKBrowsingContextController loadData:MIMEType:textEncodingName:baseURL:userData:]):
* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView loadData:MIMEType:characterEncodingName:baseURL:]):
(-[WKWebView _loadData:MIMEType:characterEncodingName:baseURL:userData:]):
* UIProcess/API/glib/WebKitWebView.cpp:
(webkit_web_view_load_html):
(webkit_web_view_load_plain_text):
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::loadData):
(WebKit::WebPageProxy::loadHTMLString): Deleted.
(WebKit::WebPageProxy::loadPlainTextString): Deleted.
* UIProcess/WebPageProxy.h:
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::loadData):
(WebKit::WebPage::loadString): Deleted.
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234954 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoCustom element doesn't invalidate its style when :defined rule starts to apply
rniwa@webkit.org [Thu, 16 Aug 2018 20:50:01 +0000 (20:50 +0000)]
Custom element doesn't invalidate its style when :defined rule starts to apply
https://bugs.webkit.org/show_bug.cgi?id=188637

Reviewed by Antti Koivisto.

LayoutTests/imported/w3c:

Rebaselined the test now that all relevant test cases pass. The only remaining test failures
are for customized builtins which we do not and shall not implement.

* web-platform-tests/custom-elements/pseudo-class-defined-expected.txt:

Source/WebCore:

Fixed the bug by invalidating the subtree when a custom element gets defined.

Test: fast/custom-elements/defined-update-style.html

* dom/Element.cpp:
(WebCore::Element::setIsDefinedCustomElement):

LayoutTests:

Added a regression test.

* fast/custom-elements/defined-update-style-expected.html: Added.
* fast/custom-elements/defined-update-style.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234953 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[ews-build] Add build steps ArchiveTestResults, UploadTestResults and ExtractTestResults
aakash_jain@apple.com [Thu, 16 Aug 2018 20:41:39 +0000 (20:41 +0000)]
[ews-build] Add build steps ArchiveTestResults, UploadTestResults and ExtractTestResults
https://bugs.webkit.org/show_bug.cgi?id=188666

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-build/steps.py:
(ArchiveTestResults): Build step to Archive the test results.
(UploadTestResults): Build step to upload the archive.
(ExtractTestResults): Build step to unzip the archive on server and generate a link.
* BuildSlaveSupport/ews-build/steps_unittest.py: Added unit-tests.
(ExpectMasterShellCommand): Copied from other similar internal code.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234952 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[ews-build] Separate queues for Builders and Testers - macOS Sierra
aakash_jain@apple.com [Thu, 16 Aug 2018 20:40:15 +0000 (20:40 +0000)]
[ews-build] Separate queues for Builders and Testers - macOS Sierra
https://bugs.webkit.org/show_bug.cgi?id=188516

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-build/config.json: Splitted macOS Sierra Release queues into builders and testers.
* BuildSlaveSupport/ews-build/factories.py: Used appropriate base classes for builders and testers.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234951 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[Curl] Test gardening
Basuke.Suzuki@sony.com [Thu, 16 Aug 2018 20:15:32 +0000 (20:15 +0000)]
[Curl] Test gardening
https://bugs.webkit.org/show_bug.cgi?id=188668

Unreviewed test gardening.

After r234863 landed, the information in the `LayoutTests/tests-options.json` is effective on WinCairo tests.
Remove [ Slow ] keywords. Also maintain xmlhttprequest results.

* platform/wincairo/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234950 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoChange the input camera in the sources & outputs example on Safari (11) with M.Way...
youenn@apple.com [Thu, 16 Aug 2018 19:50:34 +0000 (19:50 +0000)]
Change the input camera in the sources & outputs example on Safari (11) with M.Way camera
https://bugs.webkit.org/show_bug.cgi?id=187756
<rdar://problem/42332178>

Reviewed by Eric Carlson.

Covered by manual testing.

640x480x30fps is added as ideal constraints to getUserMedia if none other is added to provide consistent behavior.
Some cameras do not support 30fps as they have discrete frame rate ranges.
Before the patch, we were rejecting getUserMedia promise if the 'ideal' frameRate was not supported.
After the patch, we do not check whether frame rate is supported if it is not mandatory.
At capture time, we will still try to apply frame rate change and if not supported, the frame rate will not be set to 30 fps.

While we could surface discrete width/height/frame rate ranges, it seems best to fix the issue this way.
In the future, we might implement video interpolation for width/height/framerate at WebCore level.

* platform/mediastream/RealtimeMediaSource.cpp:
(WebCore::RealtimeMediaSource::supportsSizeAndFrameRate):
(WebCore::RealtimeMediaSource::supportsConstraint const):
(WebCore::RealtimeMediaSource::supportsConstraints):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234949 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[Win][DumpRenderTree] Remove error message for duplicated tests.
Basuke.Suzuki@sony.com [Thu, 16 Aug 2018 19:38:32 +0000 (19:38 +0000)]
[Win][DumpRenderTree] Remove error message for duplicated tests.
https://bugs.webkit.org/show_bug.cgi?id=188662

Reviewed by Per Arne Vollan.

This was added to track down an issue where the same tests appeared
to be run multiple times. This turned out to not be the case.
See https://bugs.webkit.org/show_bug.cgi?id=166760 .

* DumpRenderTree/win/DumpRenderTree.cpp:
(runTest):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234948 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[Win] The layout test editing/pasteboard/drag-and-drop-color-input-events.html is...
pvollan@apple.com [Thu, 16 Aug 2018 19:02:45 +0000 (19:02 +0000)]
[Win] The layout test editing/pasteboard/drag-and-drop-color-input-events.html is timing out.
https://bugs.webkit.org/show_bug.cgi?id=188667

Unreviewed test gardening.

* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234947 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoTransition more WKWebViewConfiguration ivars to API::PageConfiguration values
achristensen@apple.com [Thu, 16 Aug 2018 18:49:48 +0000 (18:49 +0000)]
Transition more WKWebViewConfiguration ivars to API::PageConfiguration values
https://bugs.webkit.org/show_bug.cgi?id=188665

Reviewed by Joseph Pecoraro.

APPLICATION_MANIFEST is enabled on all Cocoa platforms, so I removed some guards, too!

* UIProcess/API/APIPageConfiguration.cpp:
(API::PageConfiguration::applicationManifest const):
* UIProcess/API/APIPageConfiguration.h:
* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _initializeWithConfiguration:]):
* UIProcess/API/Cocoa/WKWebViewConfiguration.mm:
(-[WKWebViewConfiguration init]):
(-[WKWebViewConfiguration copyWithZone:]):
(-[WKWebViewConfiguration _applicationManifest]):
(-[WKWebViewConfiguration _setApplicationManifest:]):
(-[WKWebViewConfiguration _setCPULimit:]):
(-[WKWebViewConfiguration _cpuLimit]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234946 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoLayoutTest inspector/worker/debugger-pause.html sometimes times out
commit-queue@webkit.org [Thu, 16 Aug 2018 18:41:30 +0000 (18:41 +0000)]
LayoutTest inspector/worker/debugger-pause.html sometimes times out
https://bugs.webkit.org/show_bug.cgi?id=188580

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2018-08-16
Reviewed by Matt Baker.

Source/WebInspectorUI:

* UserInterface/Protocol/Target.js:
(WI.Target.prototype.get mainResource):
(WI.Target.prototype.set mainResource):
Dispatch an event when the Main Resource is set.

LayoutTests:

* inspector/worker/debugger-pause.html:
* inspector/worker/resources/worker-utilities.js: Added.
(TestPage.registerInitializer.window.awaitTargetMainResource):
(TestPage.registerInitializer):
Use a more deterministic wait for the main resource.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234945 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoPerform a microtask checkpoint before creating a custom element
rniwa@webkit.org [Thu, 16 Aug 2018 18:25:02 +0000 (18:25 +0000)]
Perform a microtask checkpoint before creating a custom element
https://bugs.webkit.org/show_bug.cgi?id=188189
<rdar://problem/42843022>

Reviewed by Geoffrey Garen.

Source/WebCore:

Fixed the bug that the HTML parser was not performing a microtask checkpoint prior to synchronously constructing
a custom element in the concept to create an element for a token:
https://html.spec.whatwg.org/multipage/parsing.html#creating-and-inserting-nodes:perform-a-microtask-checkpoint

Also added a microtask checkpoint before dispatching DOMContentLoaded to work around webkit.org/b/82931 since
scheduling a task to fire a DOMContentLoaded event in Document::finishedParsing as the HTML5 spec mandates
is a long standing bug with a lot of implications, which is completely outside the scope of this bug fix:
https://html.spec.whatwg.org/multipage/parsing.html#stop-parsing

Test: fast/custom-elements/perform-microtask-checkpoint-before-construction.html

* dom/Document.cpp:
(WebCore::Document::finishedParsing): Perform a microtask checkpoint before dispatching DOMContentLoaded here as
a workaround for webkit.org/b/82931.
* html/parser/HTMLDocumentParser.cpp:
(WebCore::HTMLDocumentParser::runScriptsForPausedTreeBuilder): Perform a microtask checkpoint here to fix the bug.

LayoutTests:

Added a W3C style testharness.js test for perfoming microtask checkpoint before constructing
a custom element synchronously.

* fast/custom-elements/perform-microtask-checkpoint-before-construction-expected.txt: Added.
* fast/custom-elements/perform-microtask-checkpoint-before-construction.html: Added.
* fast/dom/MutationObserver/parser-mutations.html: Fixed the test per new behavior in Document::finishParsing.
Because iframe loads synchronously and fires DOMContentLoaded, mutation records are now delivered twice after
iframe element is encountered in this test and before script element executes. Concatenate the mutation records
arrays to account for this behavioral change. New WebKit behavior matches that of Chrome; namely this test
fails both on Chrome Canary 70 and trunk WebKit with this patch without this fix.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234944 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoRe-introduce assertion removed in r234890
commit-queue@webkit.org [Thu, 16 Aug 2018 18:21:52 +0000 (18:21 +0000)]
Re-introduce assertion removed in r234890
https://bugs.webkit.org/show_bug.cgi?id=188611

Patch by Alex Christensen <achristensen@webkit.org> on 2018-08-16
Reviewed by Geoffrey Garen.

Source/WebCore:

The assertion was failing because Vector<String>::operator== was using memcmp,
which failed to consider different instances of equal Strings equal.  Added an
API test and this assertion, which now succeeds.

* contentextensions/ContentExtensionCompiler.cpp:
(WebCore::ContentExtensions::compileRuleList):

Source/WTF:

* wtf/text/WTFString.h:

Tools:

* TestWebKitAPI/Tests/WTF/Vector.cpp:
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234943 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoTransition more WKWebViewConfiguration ivars to API::PageConfiguration values
achristensen@apple.com [Thu, 16 Aug 2018 18:19:19 +0000 (18:19 +0000)]
Transition more WKWebViewConfiguration ivars to API::PageConfiguration values
https://bugs.webkit.org/show_bug.cgi?id=188663

Reviewed by Tim Horton.

* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _initializeWithConfiguration:]):
* UIProcess/API/Cocoa/WKWebViewConfiguration.mm:
(-[WKWebViewConfiguration init]):
(-[WKWebViewConfiguration copyWithZone:]):
(-[WKWebViewConfiguration _drawsBackground]):
(-[WKWebViewConfiguration _setDrawsBackground:]):
(-[WKWebViewConfiguration _waitsForPaintAfterViewDidMoveToWindow]):
(-[WKWebViewConfiguration _setWaitsForPaintAfterViewDidMoveToWindow:]):
(-[WKWebViewConfiguration _isControlledByAutomation]):
(-[WKWebViewConfiguration _setControlledByAutomation:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234942 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoStop using canAuthenticateAgainstProtectionSpace in modern WebKit
achristensen@apple.com [Thu, 16 Aug 2018 17:59:30 +0000 (17:59 +0000)]
Stop using canAuthenticateAgainstProtectionSpace in modern WebKit
https://bugs.webkit.org/show_bug.cgi?id=188639

Reviewed by Youenn Fablet.

canAuthenticateAgainstProtectionSpace is an unnecessary step in the authentication process.
It is leftover from when it was necessary when we used NSURLConnection, which is only used in WebKitLegacy now.
Now it's just an extra IPC roundtrip asking if we should use NSURLSessionAuthChallengeRejectProtectionSpace
or if we are going to ask the API client.  We can move this step into the C API for compatibility
with the 1 client that still uses it (not for long, see rdar://problem/43358403) and simplify and optimize
authentication.

* NetworkProcess/Downloads/PendingDownload.cpp:
(WebKit::PendingDownload::canAuthenticateAgainstProtectionSpaceAsync): Deleted.
* NetworkProcess/Downloads/PendingDownload.h:
* NetworkProcess/NetworkCORSPreflightChecker.cpp:
(WebKit::NetworkCORSPreflightChecker::didReceiveChallenge):
* NetworkProcess/NetworkLoad.cpp:
(WebKit::NetworkLoad::~NetworkLoad):
(WebKit::NetworkLoad::didReceiveChallenge):
(WebKit::NetworkLoad::completeAuthenticationChallenge): Deleted.
(WebKit::NetworkLoad::continueCanAuthenticateAgainstProtectionSpace): Deleted.
* NetworkProcess/NetworkLoad.h:
* NetworkProcess/NetworkLoadClient.h:
* NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::cancelDownload):
(WebKit::NetworkProcess::canAuthenticateAgainstProtectionSpace): Deleted.
(WebKit::NetworkProcess::continueCanAuthenticateAgainstProtectionSpace): Deleted.
* NetworkProcess/NetworkProcess.h:
* NetworkProcess/NetworkProcess.messages.in:
* NetworkProcess/NetworkResourceLoader.cpp:
(WebKit::NetworkResourceLoader::canAuthenticateAgainstProtectionSpaceAsync): Deleted.
* NetworkProcess/NetworkResourceLoader.h:
* NetworkProcess/PreconnectTask.cpp:
(WebKit::PreconnectTask::canAuthenticateAgainstProtectionSpaceAsync): Deleted.
* NetworkProcess/PreconnectTask.h:
* NetworkProcess/cache/NetworkCacheSpeculativeLoad.cpp:
(WebKit::NetworkCache::SpeculativeLoad::canAuthenticateAgainstProtectionSpaceAsync): Deleted.
* NetworkProcess/cache/NetworkCacheSpeculativeLoad.h:
* Shared/Authentication/AuthenticationManager.cpp:
(WebKit::AuthenticationManager::didReceiveAuthenticationChallenge):
* Shared/Authentication/AuthenticationManager.h:
* UIProcess/API/C/WKPage.cpp:
(WKPageSetPageNavigationClient):
* UIProcess/API/Cocoa/WKNavigationDelegatePrivate.h:
* UIProcess/Cocoa/NavigationState.h:
* UIProcess/Cocoa/NavigationState.mm:
(WebKit::NavigationState::setNavigationDelegate):
(WebKit::NavigationState::NavigationClient::didReceiveAuthenticationChallenge):
(WebKit::NavigationState::NavigationClient::canAuthenticateAgainstProtectionSpace): Deleted.
* UIProcess/Network/NetworkProcessProxy.cpp:
(WebKit::NetworkProcessProxy::canAuthenticateAgainstProtectionSpace): Deleted.
* UIProcess/Network/NetworkProcessProxy.h:
* UIProcess/Network/NetworkProcessProxy.messages.in:
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::canAuthenticateAgainstProtectionSpace): Deleted.
* UIProcess/WebPageProxy.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234941 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoAdd Experimental Feature support for SourceBuffer.changeType()
jer.noble@apple.com [Thu, 16 Aug 2018 17:57:01 +0000 (17:57 +0000)]
Add Experimental Feature support for SourceBuffer.changeType()
https://bugs.webkit.org/show_bug.cgi?id=188626
LayoutTests/imported/w3c:

Reviewed by Eric Carlson.

* resources/import-expectations.json:
* web-platform-tests/media-source/mediasource-changetype-expected.txt: Added.
* web-platform-tests/media-source/mediasource-changetype-play-expected.txt: Added.
* web-platform-tests/media-source/mediasource-changetype-play.html: Added.
* web-platform-tests/media-source/mediasource-changetype-util.js: Added.
(findSupportedChangeTypeTestTypes):
(appendBuffer):
(trimBuffered):
(trimDuration):
(runChangeTypeTest):
(mediaSourceChangeTypeTest):
* web-platform-tests/media-source/mediasource-changetype.html: Added.
* web-platform-tests/media-source/mp3/sound_5.mp3: Added.
* web-platform-tests/media-source/mp3/w3c-import.log: Added.
* web-platform-tests/media-source/w3c-import.log:
* web-platform-tests/media-source/webm/test-vp9.webm: Added.
* web-platform-tests/media-source/webm/w3c-import.log:

Source/WebCore:

Reviewed by Eric Carlson.

Tests: imported/w3c/web-platform-tests/media-source/mediasource-changetype-play.html
       imported/w3c/web-platform-tests/media-source/mediasource-changetype.html

The WICG is incubating a new proposal to allow clients to switch between streams
with different codecs within the same SourceBuffer. The proposal is being tracked
by <https://github.com/w3c/media-source/issues/155> and the proposed spec change
is available at <https://rawgit.com/wicg/media-source/codec-switching/index.html>.

Drive-by fix: One of the wpt tests uses a .mp3 file to test SourceBuffer switching.
The AVStreamDataParser does not necessarily emit every frame appended, and after a
new init segment is appended, it's possible that frames from the previous media
segment will be emitted. This causes an error condition to occur, due to an added
check in the spec for a new flag. When the SourceBuffer asks its private implementation
to reset the parser state, since there's no API on AVStreamDataParser to actually
reset the parser state, just set a flag to drop all emitted buffers until the next
initialization segment is parsed.

* Modules/mediasource/MediaSource.cpp:
(WebCore::MediaSource::contentTypeShouldGenerateTimestamps):
(WebCore::MediaSource::setDurationInternal):
(WebCore::MediaSource::addSourceBuffer):
* Modules/mediasource/MediaSource.h:
* Modules/mediasource/SourceBuffer.cpp:
(WebCore::SourceBuffer::changeType):
(WebCore::SourceBuffer::sourceBufferPrivateDidReceiveInitializationSegment):
(WebCore::SourceBuffer::validateInitializationSegment):
(WebCore::SourceBuffer::sourceBufferPrivateDidReceiveSample):
* Modules/mediasource/SourceBuffer.h:
* Modules/mediasource/SourceBuffer.idl:
* page/Settings.yaml:
* platform/graphics/SourceBufferPrivate.h:
(WebCore::SourceBufferPrivate::canSwitchToType):
* platform/graphics/avfoundation/objc/SourceBufferPrivateAVFObjC.h:
* platform/graphics/avfoundation/objc/SourceBufferPrivateAVFObjC.mm:
(WebCore::SourceBufferPrivateAVFObjC::didParseStreamDataAsAsset):
(WebCore::SourceBufferPrivateAVFObjC::processCodedFrame):
(WebCore::SourceBufferPrivateAVFObjC::resetParserState):
(WebCore::SourceBufferPrivateAVFObjC::canSwitchToType):

Source/WebKit:

Reviewed by Eric Carlson.

* Shared/WebPreferences.yaml:
* UIProcess/API/C/WKPreferences.cpp:
(WKPreferencesSetSourceBufferChangeTypeEnabled):
(WKPreferencesGetSourceBufferChangeTypeEnabled):
* UIProcess/API/C/WKPreferencesRefPrivate.h:

Source/WebKitLegacy/mac:

<rdar://problem/43356021>

Reviewed by Eric Carlson.

Add a new WebPreference property mapping to the WebCore SourceBufferChangeTypeEnabled setting.

* WebView/WebPreferenceKeysPrivate.h:
* WebView/WebPreferences.mm:
(+[WebPreferences initialize]):
(-[WebPreferences sourceBufferChangeTypeEnabled]):
(-[WebPreferences setSourceBufferChangeTypeEnabled:]):
* WebView/WebPreferencesPrivate.h:
* WebView/WebView.mm:
(-[WebView _preferencesChanged:]):

Tools:

<rdar://problem/43356021>

Reviewed by Eric Carlson.

* DumpRenderTree/mac/DumpRenderTree.mm:
(enableExperimentalFeatures):
(resetWebPreferencesToConsistentValues):
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::resetPreferencesToConsistentValues):

LayoutTests:

Reviewed by Eric Carlson.

* platform/mac/imported/w3c/web-platform-tests/media-source/mediasource-changetype-play-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234940 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoTransition more WKWebViewConfiguration ivars to API::PageConfiguration values
achristensen@apple.com [Thu, 16 Aug 2018 17:53:34 +0000 (17:53 +0000)]
Transition more WKWebViewConfiguration ivars to API::PageConfiguration values
https://bugs.webkit.org/show_bug.cgi?id=188661

Reviewed by Anders Carlsson.

* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _initializeWithConfiguration:]):
* UIProcess/API/Cocoa/WKWebViewConfiguration.mm:
(-[WKWebViewConfiguration init]):
(-[WKWebViewConfiguration copyWithZone:]):
(-[WKWebViewConfiguration _alwaysRunsAtForegroundPriority]):
(-[WKWebViewConfiguration _setAlwaysRunsAtForegroundPriority:]):
(-[WKWebViewConfiguration _initialCapitalizationEnabled]):
(-[WKWebViewConfiguration _setInitialCapitalizationEnabled:]):
(-[WKWebViewConfiguration _overrideContentSecurityPolicy]):
(-[WKWebViewConfiguration _setOverrideContentSecurityPolicy:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234939 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[Curl] Not all Cookie database files are deleted on corruption
chris.reid@sony.com [Thu, 16 Aug 2018 17:40:34 +0000 (17:40 +0000)]
[Curl] Not all Cookie database files are deleted on corruption
https://bugs.webkit.org/show_bug.cgi?id=185873

Reviewed by Fujii Hironori.

No new tests.

Delete the cookie database file and the corruption flag file when corruption is detected.
Adding a quick integrity check in case it can catch anything when opening the database.
Removing PRAGMA temp_store = MEMORY and PRAGMA journal_mode = WAL as they are set in SQLiteDatabase::open.

* platform/network/curl/CookieJarDB.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234938 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[win-ews] No such file or directory: u'/cygdrive/C/cygwin/home/buildbot/WebKit/WebKit...
pvollan@apple.com [Thu, 16 Aug 2018 17:24:46 +0000 (17:24 +0000)]
[win-ews] No such file or directory: u'/cygdrive/C/cygwin/home/buildbot/WebKit/WebKitBuild/Release/bin32/layout-test-results/pywebsocket.ws.log-out.txt'
https://bugs.webkit.org/show_bug.cgi?id=188640

Unreviewed test gardening.

For now, skip websocket tests on Windows. Currently, run-webkit-tests is failing to start the websocket server because
port 8880 is already in use. Port 8880 is being used because the python websocket server process from the previous run
did not shutdown properly. We still need to understand why this is happening, but disable websocket tests now in order
for tests to run again.

* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234933 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[Nicosia] Add Nicosia::ImageBackingTextureMapperImpl
zandobersek@gmail.com [Thu, 16 Aug 2018 16:57:35 +0000 (16:57 +0000)]
[Nicosia] Add Nicosia::ImageBackingTextureMapperImpl
https://bugs.webkit.org/show_bug.cgi?id=188550

Reviewed by Carlos Garcia Campos.

Add the Nicosia::ImageBackingTextureMapperImpl class, the
TextureMapper-specific implementation that will extend the ImageBacking
class.

This class will be used to contain and manage updates for layers that
are backed by image objects. The CoordinatedGraphicsLayer instance will
use the LayerState object to track the currently-presented image objects
and to store the latest update of these objects. The Update struct
contains information about the current visibility of the image in the
layer tree as well as the Nicosia::Buffer object containing the painted
image data that's to be presented on the screen.

During the layer flush the flushUpdate() method will move the current
update data into the pending position, from which this data will be
gathered by the composition component through the takeUpdate() method
and used to update the CoordinatedBackingStore object that's kept in
the CompositionState object on this impl class.

This will be deployed for use in the CoordinatedGraphicsLayer at a later
stage as it requires a larger rework in both CoordinatedGraphicsLayer as
well as the CoordinatedGraphicsScene classes.

The implementation itself differs from the current image backing support
in CoordinatedGraphicsLayer and CompositingCoordinator by abandoning the
idea of sharing single image backings between layers that might be
backed by identical Image objects. We thus end up trading any memory
consumption benefit in these scenarios for a much simplified
implementation. Overall this area needs more research both in figuring
out whether there's worthy improvements in grouping image backings, as
well as investigating whether it's possible to avoid additional image
rasterization by just leveraging direct pixel data via the
NativeImagePtr object.

* platform/TextureMapper.cmake:
* platform/graphics/nicosia/texmap/NicosiaImageBackingTextureMapperImpl.cpp: Added.
(Nicosia::ImageBackingTextureMapperImpl::createFactory):
(Nicosia::ImageBackingTextureMapperImpl::flushUpdate):
(Nicosia::ImageBackingTextureMapperImpl::takeUpdate):
* platform/graphics/nicosia/texmap/NicosiaImageBackingTextureMapperImpl.h: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234932 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoUnreviewed WPE gardening. Enable a few subdirectories under fast/ and
zandobersek@gmail.com [Thu, 16 Aug 2018 16:49:02 +0000 (16:49 +0000)]
Unreviewed WPE gardening. Enable a few subdirectories under fast/ and
provide platform-specific baselines as well as failure expectations
where required.

* platform/wpe/TestExpectations:
* platform/wpe/fast/box-shadow: Added 10 baselines.
* platform/wpe/fast/box-sizing: Added 4 baselines.
* platform/wpe/fast/clip: Added 23 baselines.
* platform/wpe/fast/compact: Added 3 baselines.
* platform/wpe/fast/css-generated-content: Added 43 baselines.
* platform/wpe/fast/doctypes: Added 4 baselines.
* platform/wpe/fast/flexbox: Added 26 baselines.
* platform/wpe/fast/inline: Added 32 baselines.
* platform/wpe/fast/inline-block: Added 9 baselines.
* platform/wpe/fast/invalid: Added 29 baselines.
* platform/wpe/fast/line-grid: Added 8 baselines.
* platform/wpe/fast/lists: Added 27 baselines.
* platform/wpe/fast/loader: Added 1 baseline.
* platform/wpe/fast/preloader: Added 2 baselines.
* platform/wpe/fast/replaced: Added 57 baselines.
* platform/wpe/fast/ruby: Added 26 baselines.
* platform/wpe/fast/writing-mode: Added 37 baselines.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234931 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoSupport drag-and-drop for input[type=color]
akeerthi@apple.com [Thu, 16 Aug 2018 16:44:17 +0000 (16:44 +0000)]
Support drag-and-drop for input[type=color]
https://bugs.webkit.org/show_bug.cgi?id=188464

Reviewed by Wenson Hsieh.

Source/WebCore:

input[type=color] should support native drag and drop on both macOS and iOS.

Added methods to Pasteboard and PlatformPasteboard to enable colors to be read
from and written to the pasteboard. On macOS, colors are managed through the
NSColorPboardType, whereas on iOS, colors are managed through the
'com.apple.uikit.color' identifier.

DragSourceActionColor was added to the list of DragSourceActions to identify
when a color input is being dragged.

Tests: editing/pasteboard/drag-and-drop-color-input-events.html
       editing/pasteboard/drag-and-drop-color-input.html

* page/DragActions.h:
* page/DragController.cpp:
(WebCore::DragController::dragEnteredOrUpdated):
(WebCore::isEnabledColorInput):
(WebCore::DragController::concludeEditDrag):
(WebCore::DragController::canProcessDrag):
(WebCore::DragController::draggableElement const):
(WebCore::DragController::startDrag):
* page/EventHandler.cpp:
(WebCore::EventHandler::dragHysteresisExceeded const):
(WebCore::EventHandler::handleDrag):
* page/EventHandler.h:
* page/mac/DragControllerMac.mm:
(WebCore::DragController::updateSupportedTypeIdentifiersForDragHandlingMethod const):
* platform/DragData.h:
* platform/DragImage.cpp:
(WebCore::DragImage::operator=):
* platform/DragImage.h: Added the visiblePath property to DragImage, in order for
us to be able to clip images in the UITargetedDragPreview on iOS.
* platform/DragItem.h:
(WebCore::DragItem::encode const):
(WebCore::DragItem::decode):
* platform/Pasteboard.h:
* platform/PasteboardStrategy.h:
* platform/PlatformPasteboard.h:
* platform/gtk/DragImageGtk.cpp:
(WebCore::createDragImageForColor):
* platform/gtk/PasteboardGtk.cpp:
(WebCore::Pasteboard::write):
* platform/ios/DragImageIOS.mm:
(WebCore::createDragImageForColor):
* platform/ios/PasteboardIOS.mm:
(WebCore::Pasteboard::write):
* platform/ios/PlatformPasteboardIOS.mm:
(WebCore::PlatformPasteboard::color):
(WebCore::PlatformPasteboard::setColor):
* platform/mac/DragDataMac.mm:
(WebCore::colorPasteboardType):
(WebCore::DragData::containsCompatibleContent const):
* platform/mac/DragImageMac.mm:
(WebCore::createDragImageForColor): Draw a rounded rectangle with a fill color
matching the value of the dragged color input. The rendered image is designed to
mimic the preview shown when dragging an NSColorWell.
* platform/mac/PasteboardMac.mm:
(WebCore::Pasteboard::write):
* platform/mac/PlatformPasteboardMac.mm:
(WebCore::PlatformPasteboard::setColor):
* platform/win/PasteboardWin.cpp:
(WebCore::Pasteboard::write):
* platform/wpe/PasteboardWPE.cpp:
(WebCore::Pasteboard::write):

Source/WebKit:

On iOS, the drag preview for the color input is a rounded rectangle. In order to
ensure that the corners appear transparent, the visiblePath property of the
UIDragPreviewParameters was set to match the preview's shape. This also required
the creation of an additional ArgumentCoder for Path.

When beginning the drag session, the preview should appear centered about the
color input. This is managed in createTargetedDragPreview. However, once the
preview is dragged, the preview should be at the center of the touch location.
Consequently, DragSourceActionColor was added to the list of sources that could
update the drag preview after lifting.

* Shared/WebCoreArgumentCoders.cpp:
(IPC::ArgumentCoder<Path>::decode):
* Shared/WebCoreArgumentCoders.h:
* UIProcess/Cocoa/WebPasteboardProxyCocoa.mm:
(WebKit::WebPasteboardProxy::setPasteboardColor):
* UIProcess/WebPasteboardProxy.h:
* UIProcess/WebPasteboardProxy.messages.in:
* UIProcess/ios/DragDropInteractionState.h:
* UIProcess/ios/DragDropInteractionState.mm:
(WebKit::createTargetedDragPreview):
(WebKit::shouldUseDragImageToCreatePreviewForDragSource):
(WebKit::shouldUseVisiblePathToCreatePreviewForDragSource):
(WebKit::canUpdatePreviewForActiveDragSource):
(WebKit::DragDropInteractionState::previewForDragItem const):
(WebKit::DragDropInteractionState::stageDragItem):
(WebKit::DragDropInteractionState::updatePreviewsForActiveDragSources):
* UIProcess/ios/forms/WKFormColorPicker.mm:
(-[WKColorPicker initWithView:]):
* UIProcess/mac/WebColorPickerMac.h:
* WebProcess/WebCoreSupport/WebPlatformStrategies.cpp:
(WebKit::WebPlatformStrategies::setColor):
* WebProcess/WebCoreSupport/WebPlatformStrategies.h:

Source/WebKitLegacy/mac:

* WebCoreSupport/WebPlatformStrategies.h:
* WebCoreSupport/WebPlatformStrategies.mm:
(WebPlatformStrategies::setColor):

Tools:

Added five cross-platform API tests, to verify that dragging and dropping with
color inputs changes the value of the drop target if and only if both inputs are
enabled. Also tests that the change and input events are fired when changing the
value of a color input through drag and drop.

Additionally, added a macOS-specific test to verify that dropping an item with
NSColorPboardType changes the value of the color input.

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKitCocoa/DragAndDropTests.mm:
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebKitCocoa/color-drop.html: Added.
* TestWebKitAPI/Tests/mac/DragAndDropTestsMac.mm:
(TestWebKitAPI::TEST):

LayoutTests:

Added tests to verify that dragging from one color input to another behaves as
expected in WK1. The tests consider disabled and read-only inputs in addition to
default color inputs. The WK2 implementation is tested through API tests.

* editing/pasteboard/drag-and-drop-color-input-events-expected.txt: Added.
* editing/pasteboard/drag-and-drop-color-input-events.html: Added.
* editing/pasteboard/drag-and-drop-color-input-expected.txt: Added.
* editing/pasteboard/drag-and-drop-color-input.html: Added.
* platform/ios/TestExpectations:
* platform/mac/TestExpectations:
* platform/wk2/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234930 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoAssert that calling CGSSetDenyWindowServerConnections(true) succeeds
pvollan@apple.com [Thu, 16 Aug 2018 15:53:31 +0000 (15:53 +0000)]
Assert that calling CGSSetDenyWindowServerConnections(true) succeeds
https://bugs.webkit.org/show_bug.cgi?id=188615

Reviewed by Brent Fulgham.

If the call to CGSSetDenyWindowServerConnections(true) fails, it means there are open WindowServer connections
at this point, and future WindowServer connections will not be denied. We should assert that this call succeeds.

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

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234929 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoAdd fake layout test results JSON to exercise the functionality in results.html
simon.fraser@apple.com [Thu, 16 Aug 2018 15:50:16 +0000 (15:50 +0000)]
Add fake layout test results JSON to exercise the functionality in results.html
https://bugs.webkit.org/show_bug.cgi?id=188636

Reviewed by Alexey Proskuryakov.

The results for results.html are mostly blank because the page runs without any JSON
data, so there's no testing of the functionality of the page. Add some fake JSON
data which supplies most of the different types of test result, so we're testing
at least some of the results.html script.

* fast/harness/full_results.json: Added.
* fast/harness/results-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234928 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[lFC][Floating] Add estimated margin top computation.
zalan@apple.com [Thu, 16 Aug 2018 15:43:53 +0000 (15:43 +0000)]
[lFC][Floating] Add estimated margin top computation.
https://bugs.webkit.org/show_bug.cgi?id=188619

Reviewed by Antti Koivisto.

In order to figure out whether a box should avoid a float, we need to know the final positions of both (ignore relative positioning for now).
In block formatting context the final position for a normal flow box includes
1. the static position and
2. the corresponding (non)collapsed margins.
Now the vertical margins are computed when all the descendants are finalized, because the margin values might be depending on the height of the box
(and the height might be based on the content).
So when we get to the point where we intersect the box with the float to decide if the box needs to move, we don't yet have the final vertical position.

The idea here is that as long as we don't cross the block formatting context boundary, we should be able to pre-compute the final top margin.
(if this holds true for all the cases, the estimated prefix could be removed and just use marginTop() instead.)

Covered by existing block-only tests.

* layout/blockformatting/BlockFormattingContext.cpp:
(WebCore::Layout::BlockFormattingContext::layout const):
(WebCore::Layout::BlockFormattingContext::computeEstimatedMarginTop const):
(WebCore::Layout::BlockFormattingContext::computeEstimatedMarginTopForAncestors const):
(WebCore::Layout::BlockFormattingContext::computeFloatingPosition const):
(WebCore::Layout::BlockFormattingContext::computeVerticalPositionForFloatClear const):
(WebCore::Layout::BlockFormattingContext::computeInFlowHeightAndMargin const):
* layout/blockformatting/BlockFormattingContext.h:
* layout/blockformatting/BlockMarginCollapse.cpp:
(WebCore::Layout::BlockFormattingContext::MarginCollapse::estimatedMarginTop):
* layout/displaytree/DisplayBox.cpp:
(WebCore::Display::Box::Box):
* layout/displaytree/DisplayBox.h:
(WebCore::Display::Box::setHasValidTop):
(WebCore::Display::Box::setHasValidLeft):
(WebCore::Display::Box::top const):
(WebCore::Display::Box::left const):
(WebCore::Display::Box::topLeft const):
(WebCore::Display::Box::setTopLeft):
(WebCore::Display::Box::setTop):
(WebCore::Display::Box::setLeft):
(WebCore::Display::Box::setVerticalMargin):
(WebCore::Display::Box::setEstimatedMarginTop):
(WebCore::Display::Box::estimatedMarginTop const):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234927 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[l10n] Updated Polish translation of WebKitGTK+ for 2.22
mcatanzaro@igalia.com [Thu, 16 Aug 2018 15:35:30 +0000 (15:35 +0000)]
[l10n] Updated Polish translation of WebKitGTK+ for 2.22
https://bugs.webkit.org/show_bug.cgi?id=188652

Patch by Piotr Drąg <piotrdrag@gmail.com> on 2018-08-16
Rubber-stamped by Michael Catanzaro.

* pl.po:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234926 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[LFC][BFC] Display::Box interface should reflect that padding is optional.
zalan@apple.com [Thu, 16 Aug 2018 14:40:31 +0000 (14:40 +0000)]
[LFC][BFC] Display::Box interface should reflect that padding is optional.
https://bugs.webkit.org/show_bug.cgi?id=188630

Reviewed by Antti Koivisto.

Certain type of boxes can't have paddings (see Layout::Box::isPaddingApplicable).

* layout/FormattingContext.cpp:
(WebCore::Layout::FormattingContext::computeBorderAndPadding const):
(WebCore::Layout::FormattingContext::validateGeometryConstraintsAfterLayout const):
* layout/FormattingContextGeometry.cpp:
(WebCore::Layout::FormattingContext::Geometry::outOfFlowNonReplacedVerticalGeometry):
(WebCore::Layout::FormattingContext::Geometry::outOfFlowNonReplacedHorizontalGeometry):
(WebCore::Layout::FormattingContext::Geometry::outOfFlowReplacedVerticalGeometry):
(WebCore::Layout::FormattingContext::Geometry::outOfFlowReplacedHorizontalGeometry):
* layout/blockformatting/BlockFormattingContextGeometry.cpp:
(WebCore::Layout::BlockFormattingContext::Geometry::inFlowNonReplacedHeightAndMargin):
(WebCore::Layout::BlockFormattingContext::Geometry::inFlowNonReplacedWidthAndMargin):
* layout/displaytree/DisplayBox.h:
(WebCore::Display::Box::width const):
(WebCore::Display::Box::height const):
(WebCore::Display::Box::contentBoxTop const):
(WebCore::Display::Box::contentBoxLeft const):
(WebCore::Display::Box::setPadding):
(WebCore::Display::Box::paddingTop const):
(WebCore::Display::Box::paddingLeft const):
(WebCore::Display::Box::paddingBottom const):
(WebCore::Display::Box::paddingRight const):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234925 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[LFC] Add showLayoutTree() that does not require LayoutContext.
zalan@apple.com [Thu, 16 Aug 2018 14:35:24 +0000 (14:35 +0000)]
[LFC] Add showLayoutTree() that does not require LayoutContext.
https://bugs.webkit.org/show_bug.cgi?id=188631

Reviewed by Antti Koivisto.

* layout/layouttree/LayoutBox.cpp:
(WebCore::Layout::Box::formattingContextRoot const):
(WebCore::Layout::Box::initialContainingBlock const):
* layout/layouttree/LayoutBox.h:
* layout/layouttree/LayoutTreeBuilder.cpp:
(WebCore::Layout::outputLayoutTree):
(WebCore::Layout::showLayoutTree):
(WebCore::Layout::TreeBuilder::showLayoutTree): Deleted.
* layout/layouttree/LayoutTreeBuilder.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234924 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[LFC] Tree builder should construct block and inline containers based on the display...
zalan@apple.com [Thu, 16 Aug 2018 14:32:46 +0000 (14:32 +0000)]
[LFC] Tree builder should construct block and inline containers based on the display type.
https://bugs.webkit.org/show_bug.cgi?id=188632

Reviewed by Antti Koivisto.

Inline elements can also construct RenderBlock renderers (inline-bloc etc), so use the display type instead when deciding what type of container to construct.

* layout/layouttree/LayoutTreeBuilder.cpp:
(WebCore::Layout::TreeBuilder::createSubTree):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234923 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[LFC] Bail out of subtree verification when trees are out of sync.
zalan@apple.com [Thu, 16 Aug 2018 14:28:34 +0000 (14:28 +0000)]
[LFC] Bail out of subtree verification when trees are out of sync.
https://bugs.webkit.org/show_bug.cgi?id=188633

Reviewed by Antti Koivisto.

There's no point of trying to match geometry when subtrees are out of sync.
They get out of sync when the render tree has unsupported boxes (see LayoutTreeBuilder).

* layout/Verification.cpp:
(WebCore::Layout::verifyAndOutputSubtree):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234922 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoUnreviewed, WPE build fix after r234920.
philn@webkit.org [Thu, 16 Aug 2018 14:23:35 +0000 (14:23 +0000)]
Unreviewed, WPE build fix after r234920.

* UIProcess/API/wpe/PageClientImpl.cpp:
(WebKit::PageClientImpl::isViewWindowActive):
(WebKit::PageClientImpl::isViewFocused):
(WebKit::PageClientImpl::isViewVisible):
(WebKit::PageClientImpl::isViewInWindow):
* UIProcess/API/wpe/WPEView.cpp:
(WKWPE::View::setViewState):
* UIProcess/API/wpe/WPEView.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234921 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoUse OptionSet for ActivityState::Flags
antti@apple.com [Thu, 16 Aug 2018 12:49:01 +0000 (12:49 +0000)]
Use OptionSet for ActivityState::Flags
https://bugs.webkit.org/show_bug.cgi?id=188554

Reviewed by Brent Fulgham.

Source/WebCore:

More typesafe flags.

* Modules/geolocation/GeolocationController.cpp:
(WebCore::GeolocationController::activityStateDidChange):
* Modules/geolocation/GeolocationController.h:
* html/canvas/WebGLRenderingContextBase.cpp:
(WebCore::WebGLRenderingContextBase::activityStateDidChange):
* html/canvas/WebGLRenderingContextBase.h:
* page/ActivityState.cpp:
(WebCore::operator<<):
(WebCore::activityStateFlagsToString): Deleted.
* page/ActivityState.h:
(WebCore::ActivityState::allFlags):
* page/ActivityStateChangeObserver.h:
* page/FocusController.cpp:
(WebCore::FocusController::FocusController):
(WebCore::FocusController::setFocused):
(WebCore::FocusController::setActivityState):
(WebCore::FocusController::setActive):
* page/FocusController.h:
(WebCore::FocusController::isActive const):
(WebCore::FocusController::isFocused const):
(WebCore::FocusController::contentIsVisible const):
* page/Page.cpp:
(WebCore::pageInitialActivityState):
(WebCore::Page::Page):
(WebCore::Page::setIsInWindow):
(WebCore::Page::updateTimerThrottlingState):
(WebCore::Page::setActivityState):
(WebCore::Page::isVisibleAndActive const):
(WebCore::Page::isWindowActive const):
(WebCore::Page::setIsVisible):
* page/Page.h:
(WebCore::Page::activityState const):
(WebCore::Page::isVisible const):
(WebCore::Page::isInWindow const):
* page/PerformanceMonitor.cpp:
(WebCore::activityStateForCPUSampling):
(WebCore::PerformanceMonitor::activityStateChanged):
* page/PerformanceMonitor.h:
* platform/text/cocoa: Added.

Source/WebKit:

* Shared/WebPageCreationParameters.h:
* UIProcess/Cocoa/WebViewImpl.mm:
(WebKit::WebViewImpl::windowDidOrderOffScreen):
(WebKit::WebViewImpl::windowDidOrderOnScreen):
(WebKit::WebViewImpl::viewDidMoveToWindow):
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::updateActivityState):
(WebKit::WebPageProxy::activityStateDidChange):
(WebKit::WebPageProxy::dispatchActivityStateChange):
(WebKit::WebPageProxy::setMuted):
(WebKit::WebPageProxy::isPlayingMediaDidChange):
* UIProcess/WebPageProxy.h:
(WebKit::WebPageProxy::isInWindow const):
(WebKit::WebPageProxy::isViewVisible const):
(WebKit::WebPageProxy::isViewFocused const):
(WebKit::WebPageProxy::isViewWindowActive const):
* WebProcess/Plugins/PluginView.cpp:
(WebKit::PluginView::activityStateDidChange):
* WebProcess/Plugins/PluginView.h:
* WebProcess/WebPage/DrawingArea.h:
(WebKit::DrawingArea::activityStateDidChange):
* WebProcess/WebPage/RemoteLayerTree/RemoteLayerTreeDrawingArea.h:
* WebProcess/WebPage/RemoteLayerTree/RemoteLayerTreeDrawingArea.mm:
(WebKit::RemoteLayerTreeDrawingArea::activityStateDidChange):
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::updateThrottleState):
(WebKit::WebPage::updateIsInWindow):
(WebKit::WebPage::visibilityDidChange):
(WebKit::WebPage::setActivityState):
* WebProcess/WebPage/WebPage.h:
(WebKit::WebPage::isVisible const):
(WebKit::WebPage::isVisibleOrOccluded const):
* WebProcess/WebPage/WebPage.messages.in:
* WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.h:
* WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.mm:
(WebKit::TiledCoreAnimationDrawingArea::activityStateDidChange):
* WebProcess/WebProcess.cpp:
(WebKit::WebProcess::pageActivityStateDidChange):
* WebProcess/WebProcess.h:

Source/WTF:

* wtf/OptionSet.h:
(WTF::OptionSet::operator^):

Add xor operator, useful for finding changes between sets.

Tools:

* TestWebKitAPI/Tests/WTF/OptionSet.cpp:
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234920 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[YARR] Align allocation size in BumpPointerAllocator with sizeof(void*)
yusukesuzuki@slowstart.org [Thu, 16 Aug 2018 09:41:36 +0000 (09:41 +0000)]
[YARR] Align allocation size in BumpPointerAllocator with sizeof(void*)
https://bugs.webkit.org/show_bug.cgi?id=188571

Reviewed by Saam Barati.

UBSan finds YarrInterpreter performs misaligned accesses. This is because YarrInterpreter
allocates DisjunctionContext and ParenthesesDisjunctionContext from BumpPointerAllocator
without considering alignment of them. This patch adds DisjunctionContext::allocationSize
and ParenthesesDisjunctionContext::allocationSize to calculate allocation sizes for them.
The size is always rounded to `sizeof(void*)` so that these classes are always allocated
with `sizeof(void*)` alignment. We also ensure the alignments of both classes are less
than or equal to `sizeof(void*)` by `static_assert`.

* yarr/YarrInterpreter.cpp:
(JSC::Yarr::Interpreter::DisjunctionContext::allocationSize):
(JSC::Yarr::Interpreter::allocDisjunctionContext):
(JSC::Yarr::Interpreter::ParenthesesDisjunctionContext::ParenthesesDisjunctionContext):
(JSC::Yarr::Interpreter::ParenthesesDisjunctionContext::getDisjunctionContext):
(JSC::Yarr::Interpreter::ParenthesesDisjunctionContext::allocationSize):
(JSC::Yarr::Interpreter::allocParenthesesDisjunctionContext):
(JSC::Yarr::Interpreter::Interpreter):
(JSC::Yarr::Interpreter::DisjunctionContext::DisjunctionContext): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234916 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[GTK] MiniBrowser: web view doesn't get the focus when new window is created
carlosgc@webkit.org [Thu, 16 Aug 2018 08:15:54 +0000 (08:15 +0000)]
[GTK] MiniBrowser: web view doesn't get the focus when new window is created
https://bugs.webkit.org/show_bug.cgi?id=188546

Reviewed by Michael Catanzaro.

Tools:

When a new window is created, the first widget of the windo is getting the focus, the toolbar in this case. We
always want the WebView to grab the focus. In case of new empty tab, we want to the URL bar to be focused
instead. This was causing failures in WebDriver tests that create a popup window with an input element that is
filled with send keys command. The input element is marked as focused in the DOM, but the actual focused widget
is the toolbar, so key events are sent to the toolbar instead of the web view.

* MiniBrowser/gtk/BrowserWindow.c:
(webViewCreate):
(newTabCallback):
(openPrivateWindow):
(browser_window_get_or_create_web_view_for_automation):

WebDriverTests:

Remove expectations of the tests that pass now.

* TestExpectations.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234915 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoUnreviewed. Fix WebDriver tests after r234839.
carlosgc@webkit.org [Thu, 16 Aug 2018 08:14:41 +0000 (08:14 +0000)]
Unreviewed. Fix WebDriver tests after r234839.

New pytest requires to autoinstall attrs too.

* Scripts/webkitpy/thirdparty/__init__.py:
(AutoinstallImportHook._install_pytest):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234914 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agobmalloc: Coverity scan issues
tpopela@redhat.com [Thu, 16 Aug 2018 07:41:46 +0000 (07:41 +0000)]
bmalloc: Coverity scan issues
https://bugs.webkit.org/show_bug.cgi?id=186763

Reviewed by Saam Barati.

* bmalloc/DebugHeap.h: Initialize the m_pageSize variable.
* bmalloc/IsoTLS.cpp:
(bmalloc::IsoTLS::ensureEntries): Check the return value of
pthread_key_create return().
* bmalloc/VMAllocate.h:
(bmalloc::vmPageSize): Correctly check the return value of sysconf().

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234913 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoNSURLAuthenticationMethodOAuth challenges are surfaced to clients in -didReceiveAuthe...
commit-queue@webkit.org [Thu, 16 Aug 2018 05:38:39 +0000 (05:38 +0000)]
NSURLAuthenticationMethodOAuth challenges are surfaced to clients in -didReceiveAuthenticationChallenge as NSURLAuthenticationMethodDefault
https://bugs.webkit.org/show_bug.cgi?id=186870
Source/WebCore:

<rdar://problem/41314410>

Patch by Ansh Shukla <ansh_shukla@apple.com> on 2018-08-15
Reviewed by Alex Christensen.

Add the ProtectionSpaceAuthenticationSchemeOAuth type.

* platform/network/ProtectionSpaceBase.cpp:
(WebCore::ProtectionSpaceBase::isPasswordBased const): Return yes because the oauth challenge
expects a token in return.
* platform/network/ProtectionSpaceBase.h:
* platform/network/cocoa/ProtectionSpaceCocoa.mm:
(WebCore::scheme):
(WebCore::ProtectionSpace::nsSpace const):

Source/WebCore/PAL:

<rdar://problem/41314410>

Patch by Ansh Shukla <ansh_shukla@apple.com> on 2018-08-15
Reviewed by Alex Christensen.

* pal/spi/cf/CFNetworkSPI.h: Declare OAuth string when not building against the
internal SDK.

Source/WebKit:

<rdar://problem/41314410>

Patch by Ansh Shukla <ansh_shukla@apple.com> on 2018-08-15
Reviewed by Alex Christensen.

Correctly expose the OAuth protection space type in API.

* UIProcess/API/C/WKAPICast.h:
(WebKit::toAPI):
* UIProcess/API/C/WKProtectionSpaceTypes.h:

Tools:

<rdar://problem/41314410>

Patch by Ansh Shukla <ansh_shukla@apple.com> on 2018-08-15
Reviewed by Alex Christensen.

* WebKitTestRunner/TestController.cpp:
(WTR::toString):
(WTR::TestController::canAuthenticateAgainstProtectionSpace): Expose type of authentication challenge so we can test OAuth.
(WTR::TestController::didReceiveAuthenticationChallenge):

LayoutTests:

Patch by Ansh Shukla <ansh_shukla@apple.com> on 2018-08-15
Reviewed by Alex Christensen.

Ensure the exposed authentication type to clients is OAuth.

* http/tests/loading/oauth-expected.txt: Added.
* http/tests/loading/oauth.html: Added.
* http/tests/loading/resources/oauth-subresource.php: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234912 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoWe should cache the compiled sandbox profile in a data vault
commit-queue@webkit.org [Thu, 16 Aug 2018 03:34:01 +0000 (03:34 +0000)]
We should cache the compiled sandbox profile in a data vault
https://bugs.webkit.org/show_bug.cgi?id=184991

Patch by Ben Richards <benton_richards@apple.com> on 2018-08-15
Reviewed by Ryosuke Niwa.

Source/WebCore:

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

* platform/FileHandle.cpp:
(WebCore::FileHandle::FileHandle):
(WebCore::FileHandle::open):
(WebCore::FileHandle::close):
* platform/FileHandle.h:
* platform/FileSystem.h:
* platform/cocoa/FileSystemCocoa.mm:
(WebCore::FileSystem::deleteNonEmptyDirectory):

Source/WebKit:

This patch changes a few things (note: data vaults and sandbox entitlements are only used in internal builds):
(1) Instead of compiling a sandbox every time a process is launched, processes now look for a cached sandbox
    in a process specific data vault on macOS platforms. (ChildProcessMac.mm)
(2) If a valid cached sandbox is not found, a process will create the data vault (or ensure that it exists),
    compile a sandbox, and cache it.
(3) In order to create process specific data vaults, each process now has their own <process name>-OSX-sandbox.entitlements
    file which contains an entitlement with a process specific "storage class" which ensures that each process
    can only ever access its own data vault. (See the article on confluence "Data Vaults and Restricted Files" for more info)
(4) The sandbox entitlements file for the Network and WebContent services are loaded dynamically
    through Scripts/<process name>-process-entitlements.sh which is triggered in a new build phase for each service.
    The Storage process sandbox entitlements are loaded directly in Configurations/StorageService.xcconfig.
    The reason that the sandbox entitlements are applied dynamically is so that these sandbox entitlements
    are only applied when WK_USE_RESTRICTED_ENTITLEMENTS is YES. This means that open source builds will still work.

* Configurations/Network-OSX-sandbox.entitlements: Added.
* Configurations/Storage-OSX-sandbox.entitlements: Added.
* Configurations/StorageService.xcconfig:
* Configurations/WebContent-OSX-sandbox.entitlements: Added.
* Configurations/WebKit.xcconfig:
* NetworkProcess/NetworkProcess.h:
* PluginProcess/PluginProcess.h:
* Scripts/process-network-sandbox-entitlements.sh: Added.
* Scripts/process-webcontent-sandbox-entitlements.sh: Added.
* Shared/ChildProcess.h:
* Shared/EntryPointUtilities/mac/XPCService/XPCServiceEntryPoint.h:
(WebKit::XPCServiceInitializer):
* Shared/SandboxInitializationParameters.h:
(WebKit::SandboxInitializationParameters::setOverrideSandboxProfilePath):
(WebKit::SandboxInitializationParameters::overrideSandboxProfilePath const):
(WebKit::SandboxInitializationParameters::setSandboxProfile):
(WebKit::SandboxInitializationParameters::sandboxProfile const):
(): Deleted.
* Shared/mac/ChildProcessMac.mm:
(WebKit::SandboxProfileDeleter::operator()):
(WebKit::SandboxParametersDeleter::operator()):
(WebKit::SandboxInfo::SandboxInfo):
(WebKit::fileContents):
(WebKit::processStorageClass):
(WebKit::setAndSerializeSandboxParameters):
(WebKit::sandboxDataVaultParentDirectory):
(WebKit::sandboxDirectory):
(WebKit::sandboxFilePath):
(WebKit::ensureSandboxCacheDirectory):
(WebKit::writeSandboxDataToCacheFile):
(WebKit::compileAndCacheSandboxProfile):
(WebKit::tryApplyCachedSandbox):
(WebKit::webKit2Bundle):
(WebKit::getSandboxProfileOrProfilePath):
(WebKit::compileAndApplySandboxSlowCase):
(WebKit::applySandbox):
(WebKit::initializeSandboxParameters):
(WebKit::ChildProcess::initializeSandbox):
* Shared/mac/SandboxInitialiationParametersMac.mm:
(WebKit::SandboxInitializationParameters::SandboxInitializationParameters):
* StorageProcess/StorageProcess.h:
* WebKit.xcodeproj/project.pbxproj:
* WebProcess/WebProcess.h:

Source/WTF:

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

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

Tools:

Added trace points for sandbox initialization

* Tracing/SystemTracePoints.plist:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234910 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[WinCairo] Unreviewed build fix after r234896.
ross.kirsling@sony.com [Thu, 16 Aug 2018 00:04:48 +0000 (00:04 +0000)]
[WinCairo] Unreviewed build fix after r234896.

* NetworkProcess/curl/NetworkDataTaskCurl.cpp:
(WebKit::NetworkDataTaskCurl::tryHttpAuthentication):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234908 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoCan't share an app on AppStore to WeChat due to a release assert
rniwa@webkit.org [Wed, 15 Aug 2018 23:37:31 +0000 (23:37 +0000)]
Can't share an app on AppStore to WeChat due to a release assert
https://bugs.webkit.org/show_bug.cgi?id=188621
<rdar://problem/43343976>

Reviewed by Geoffrey Garen.

Disable the thread safety check when the app is not linked on or after iOS 12 since this release assert
is getting hit by third party applications on iOS in UI process.

Source/WebCore:

* platform/Timer.cpp:
(WebCore::shouldSuppressThreadSafetyCheck): Added a SDK check.

Source/WebKit:

* UIProcess/Cocoa/VersionChecks.h:
(WebKit::SDKVersion::FirstWithMainThreadReleaseAssertionInWebPageProxy): Added. It's iOS 12 or macOS 10.14 Mojave.
* UIProcess/WebProcessProxy.cpp:
(WebKit::isMainThreadOrCheckDisabled): Added. Returns true whether when we're in the main thread or if the app
is not linked on or after iOS 12 or macOS 10.14 Mojave.
(WebKit::globalPageMap):
(WebKit::m_isInPrewarmedPool):
(WebKit::WebProcessProxy::~WebProcessProxy):
(WebKit::WebProcessProxy::shutDown):
(WebKit::WebProcessProxy::deleteWebsiteDataForTopPrivatelyControlledDomainsInAllPersistentDataStores):
(WebKit::WebProcessProxy::topPrivatelyControlledDomainsWithWebsiteData):
(WebKit::WebProcessProxy::didFinishLaunching):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234905 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoRemove WKNavigationDelegatePrivate's canAuthenticateAgainstProtectionSpace
achristensen@apple.com [Wed, 15 Aug 2018 22:48:05 +0000 (22:48 +0000)]
Remove WKNavigationDelegatePrivate's canAuthenticateAgainstProtectionSpace
https://bugs.webkit.org/show_bug.cgi?id=188622

Reviewed by Timothy Hatcher.

It's been deprecated for a release now, nobody uses it, and it's a concept from NSURLConnection, which we don't use any more in WebKit2.

* UIProcess/API/Cocoa/WKNavigationDelegatePrivate.h:
* UIProcess/Cocoa/NavigationState.h:
* UIProcess/Cocoa/NavigationState.mm:
(WebKit::NavigationState::setNavigationDelegate):
(WebKit::NavigationState::NavigationClient::canAuthenticateAgainstProtectionSpace):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234903 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[Curl] Implement default cookie path handling correctly as outlined in RFC6265.
chris.reid@sony.com [Wed, 15 Aug 2018 22:03:29 +0000 (22:03 +0000)]
[Curl] Implement default cookie path handling correctly as outlined in RFC6265.
https://bugs.webkit.org/show_bug.cgi?id=188609

Reviewed by Alex Christensen.

Curl implementation of default cookie path was wrong so that some cookies cannot be accessible.
It should be generated as outlined in: https://tools.ietf.org/html/rfc6265#section-5.1.4

Source/WebCore:

Tests: http/tests/cookies/http-get-cookie-set-in-js.html

* platform/network/curl/CookieJarDB.cpp:
(WebCore::CookieJarDB::setCookie):
* platform/network/curl/CookieUtil.cpp:
(WebCore::CookieUtil::defaultPathForURL):
* platform/network/curl/CookieUtil.h:

LayoutTests:

* platform/wincairo/TestExpectations:
  - http/tests/cookies/http-get-cookie-set-in-js.html [ Pass ]

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234901 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoCrashes in Quip under _dictionaryPopupInfoForRange, in setObject:forKey:
timothy_horton@apple.com [Wed, 15 Aug 2018 20:59:57 +0000 (20:59 +0000)]
Crashes in Quip under _dictionaryPopupInfoForRange, in setObject:forKey:
https://bugs.webkit.org/show_bug.cgi?id=188569
<rdar://problem/34201095>

Reviewed by Megan Gardner.

Source/WebKit:

* WebProcess/WebPage/mac/WebPageMac.mm:
(WebKit::WebPage::dictionaryPopupInfoForRange):
Speculative fix; the crashes indicate font is null, but we just checked it,
so it must be getting made null by convertFont:toSize:. Check again!

Source/WebKitLegacy/mac:

* WebView/WebImmediateActionController.mm:
(+[WebImmediateActionController _dictionaryPopupInfoForRange:inFrame:withLookupOptions:indicatorOptions:transition:]):
Speculative fix; the crashes indicate font is null, but we just checked it,
so it must be getting made null by convertFont:toSize:. Check again!

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234900 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoDisable IntersectionObserver tests on Windows for now
ajuma@chromium.org [Wed, 15 Aug 2018 20:45:41 +0000 (20:45 +0000)]
Disable IntersectionObserver tests on Windows for now
https://bugs.webkit.org/show_bug.cgi?id=188613

Unreviewed test gardening.

The experimental feature for IntersectionObserver is off by default, and
Windows WebKit API would have to change to enable it, so just skip the
tests for now.

* platform/win/TestExpectations:
* platform/wincairo/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234899 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[Datalist] Add button to TextFieldInputs with a datalist
akeerthi@apple.com [Wed, 15 Aug 2018 20:42:26 +0000 (20:42 +0000)]
[Datalist] Add button to TextFieldInputs with a datalist
https://bugs.webkit.org/show_bug.cgi?id=187741

Reviewed by Tim Horton.

Source/WebCore:

TextFieldInputs that have an associated datalist element should be drawn as
combo boxes. However, we cannot use NSComboBox for this control, as NSComboBox
is not height-resizable. Furthermore, the input should also be able to contain
additional elements, such as the stepper for type=number and the cancel button
for type=search. For these reasons, we draw a button at the end of the input,
mimicking appearance of a combo box.

The list-button -webkit-appearance value was added to display the new button.

Tests: fast/forms/datalist/datalist-searchinput-appearance.html
       fast/forms/datalist/datalist-textinput-appearance.html

* Resources/ListButtonArrow.png: Added.
* Resources/ListButtonArrow@2x.png: Added.
* WebCore.xcodeproj/project.pbxproj:
* css/CSSPrimitiveValueMappings.h:
(WebCore::CSSPrimitiveValue::CSSPrimitiveValue):
* css/CSSProperties.json:
* css/CSSValueKeywords.in:
* css/html.css:
(input::-webkit-list-button):
* html/HTMLInputElement.cpp:
(WebCore::HTMLInputElement::dataListButtonElement const):
* html/HTMLInputElement.h:
* html/InputType.h:
(WebCore::InputType::dataListButtonElement const):
* html/TextFieldInputType.cpp:
(WebCore::TextFieldInputType::needsContainer const):
(WebCore::TextFieldInputType::createShadowSubtree):
(WebCore::TextFieldInputType::destroyShadowSubtree):
(WebCore::TextFieldInputType::listAttributeTargetChanged):
(WebCore::TextFieldInputType::dataListButtonElement const):
(WebCore::TextFieldInputType::dataListButtonElementWasClicked):
(WebCore::TextFieldInputType::didCloseSuggestions):
* html/TextFieldInputType.h:
* html/shadow/DataListButtonElement.cpp: Added.
(WebCore::DataListButtonElement::create):
(WebCore::DataListButtonElement::DataListButtonElement):
(WebCore::DataListButtonElement::~DataListButtonElement):
(WebCore::DataListButtonElement::defaultEventHandler):
* html/shadow/DataListButtonElement.h: Added.
* platform/ThemeTypes.h:
* rendering/RenderTheme.cpp:
(WebCore::RenderTheme::adjustStyle):
(WebCore::RenderTheme::adjustListButtonStyle const):
* rendering/RenderTheme.h:
* rendering/RenderThemeMac.h:
* rendering/RenderThemeMac.mm:
(-[WebListButtonCell drawWithFrame:inView:]):
(WebCore::RenderThemeMac::paintListButtonForInput):
(WebCore::RenderThemeMac::adjustListButtonStyle const):
(WebCore::RenderThemeMac::paintTextField):
(WebCore::RenderThemeMac::paintSearchField):
(WebCore::RenderThemeMac::paintSearchFieldCancelButton):
(WebCore::RenderThemeMac::listButton const):

Source/WebCore/PAL:

* pal/spi/cocoa/NSColorSPI.h: Added NSColorGetUserAccentColor().

Source/WebInspectorUI:

Add keyword completion for 'list-button'.

* UserInterface/External/CodeMirror/css.js:
* UserInterface/Models/CSSKeywordCompletions.js:

LayoutTests:

Added tests to verify appearance of TextFieldInputs with a datalist.

* fast/forms/datalist/datalist-searchinput-appearance.html: Added.
* fast/forms/datalist/datalist-textinput-appearance.html: Added.
* platform/ios/TestExpectations:
* platform/mac/fast/forms/datalist/datalist-searchinput-appearance-expected.png: Added.
* platform/mac/fast/forms/datalist/datalist-searchinput-appearance-expected.txt: Added.
* platform/mac/fast/forms/datalist/datalist-textinput-appearance-expected.png: Added.
* platform/mac/fast/forms/datalist/datalist-textinput-appearance-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234898 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoUnreviewed, rolling out r234870.
ryanhaddad@apple.com [Wed, 15 Aug 2018 20:22:12 +0000 (20:22 +0000)]
Unreviewed, rolling out r234870.

The test introduced with this change is a flaky failure.

Reverted changeset:

"NSURLAuthenticationMethodOAuth challenges are surfaced to
clients in -didReceiveAuthenticationChallenge as
NSURLAuthenticationMethodDefault"
https://bugs.webkit.org/show_bug.cgi?id=186870
https://trac.webkit.org/changeset/234870

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234897 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoNetworkCORSPreflightChecker should proceed in case of ProtectionSpaceAuthenticationSc...
achristensen@apple.com [Wed, 15 Aug 2018 20:15:57 +0000 (20:15 +0000)]
NetworkCORSPreflightChecker should proceed in case of ProtectionSpaceAuthenticationSchemeServerTrustEvaluationRequested even though the WebKit app is not implementing the didReceiveAuthenticationChallenge/didReceiveAuthenticationChallengeInFrame callback
https://bugs.webkit.org/show_bug.cgi?id=188592
<rdar://problem/43210331>

Reviewed by Youenn Fablet.

Do a canAuthenticateAgainstProtectionSpace check in NetworkCORSPreflightChecker like we do in NetworkLoad.
Use CompletionHandlers to make the now 3 different canAuthenticateAgainstProtectionSpace checks look the same from the NetworkProcess.

* NetworkProcess/NetworkCORSPreflightChecker.cpp:
(WebKit::NetworkCORSPreflightChecker::didReceiveChallenge):
* NetworkProcess/NetworkCORSPreflightChecker.h:
* NetworkProcess/NetworkDataTask.h:
* NetworkProcess/NetworkLoad.cpp:
(WebKit::NetworkLoad::didReceiveChallenge):
* NetworkProcess/NetworkLoad.h:
* NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::canAuthenticateAgainstProtectionSpace):
(WebKit::NetworkProcess::continueCanAuthenticateAgainstProtectionSpace):
* NetworkProcess/NetworkProcess.h:
* NetworkProcess/NetworkResourceLoader.cpp:
(WebKit::NetworkResourceLoader::canAuthenticateAgainstProtectionSpaceAsync):
(WebKit::NetworkResourceLoader::continueCanAuthenticateAgainstProtectionSpace): Deleted.
* NetworkProcess/NetworkResourceLoader.h:
* NetworkProcess/PingLoad.cpp:
(WebKit::PingLoad::didReceiveChallenge):
* NetworkProcess/PingLoad.h:
* NetworkProcess/PreconnectTask.cpp:
(WebKit::PreconnectTask::canAuthenticateAgainstProtectionSpaceAsync):
(WebKit::PreconnectTask::continueCanAuthenticateAgainstProtectionSpace): Deleted.
* NetworkProcess/PreconnectTask.h:
* NetworkProcess/cocoa/NetworkDataTaskCocoa.h:
* NetworkProcess/cocoa/NetworkDataTaskCocoa.mm:
(WebKit::NetworkDataTaskCocoa::didReceiveChallenge):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234896 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[Curl] Test gardening
Basuke.Suzuki@sony.com [Wed, 15 Aug 2018 19:44:27 +0000 (19:44 +0000)]
[Curl] Test gardening
https://bugs.webkit.org/show_bug.cgi?id=188610

Unreviewed test gardening.

Unskip http/tests/cookies.

* platform/wincairo/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234895 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoRemove evernote hacks
keith_miller@apple.com [Wed, 15 Aug 2018 19:33:54 +0000 (19:33 +0000)]
Remove evernote hacks
https://bugs.webkit.org/show_bug.cgi?id=188591

Reviewed by Joseph Pecoraro.

The hack was added in 2012 and the evernote app seems to work now.
It's probably not needed anymore.

* API/JSValueRef.cpp:
(JSValueUnprotect):
(evernoteHackNeeded): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234894 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoconnectedCallback is invoked by the HTML parser after child nodes had been inserted
rniwa@webkit.org [Wed, 15 Aug 2018 18:59:19 +0000 (18:59 +0000)]
connectedCallback is invoked by the HTML parser after child nodes had been inserted
https://bugs.webkit.org/show_bug.cgi?id=183931
<rdar://problem/38843548>

Reviewed by Alex Christensen.

LayoutTests/imported/w3c:

Rebaselined the test now that all test cases pass.

* web-platform-tests/custom-elements/parser/parser-sets-attributes-and-children-expected.txt:

Source/WebCore:

Invoke the custom element reactions after constructing and inserting a custom element as specifed in step 3.3 of:
https://html.spec.whatwg.org/multipage/parsing.html#insert-a-foreign-element

The bug here was that HTMLConstructionSite::insertCustomElement uses attachLater so that even though the task
to insert the custom element was created, it didn't get executed until after CustomElementReactionStack in
HTMLDocumentParser::runScriptsForPausedTreeBuilder had been popped off of the stack.

Test: imported/w3c/web-platform-tests/custom-elements/parser/parser-sets-attributes-and-children.html

* html/parser/HTMLConstructionSite.cpp:
(WebCore::HTMLConstructionSite::insertCustomElement): Fixed the bug by manually executing the scheduled tasks.
This will enqueue any custom element reactions while CustomElementReactionStack in runScriptsForPausedTreeBuilder
is still in the stack.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234893 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[Curl] Don't send Content-Type header for POST request when body is null.
Basuke.Suzuki@sony.com [Wed, 15 Aug 2018 18:50:54 +0000 (18:50 +0000)]
[Curl] Don't send Content-Type header for POST request when body is null.
https://bugs.webkit.org/show_bug.cgi?id=188588

Reviewed by Youenn Fablet.

Source/WebCore:

The Content-Type header was sent by libcurl automatically. Suppress that behavior
when body is null.

Tests: http/tests/xmlhttprequest/methods.html

* platform/network/curl/CurlRequest.cpp:
(WebCore::CurlRequest::setupPOST):

LayoutTests:

* platform/wincairo/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234892 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoUnreviewed, change my status to be a WebKit reviewer
megan_gardner@apple.com [Wed, 15 Aug 2018 17:59:16 +0000 (17:59 +0000)]
Unreviewed, change my status to be a WebKit reviewer
and added expertise.

* Scripts/webkitpy/common/config/contributors.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234891 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoRemove failing assertion introduced in r234873
achristensen@apple.com [Wed, 15 Aug 2018 17:48:07 +0000 (17:48 +0000)]
Remove failing assertion introduced in r234873
https://bugs.webkit.org/show_bug.cgi?id=188581

* contentextensions/ContentExtensionCompiler.cpp:
(WebCore::ContentExtensions::compileRuleList):
The assertion is correct but failing because VectorTraits<String> is incorrect.
I'll re-add it and fix VectorTraits<String> in a separate patch.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234890 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoWeb Inspector: REGRESSION(r?): the probe sidebar doesn't show up when adding probes
drousso@apple.com [Wed, 15 Aug 2018 17:29:14 +0000 (17:29 +0000)]
Web Inspector: REGRESSION(r?): the probe sidebar doesn't show up when adding probes
https://bugs.webkit.org/show_bug.cgi?id=188594

Reviewed by Brian Burg.

* UserInterface/Views/ProbeDetailsSidebarPanel.js:
(WI.ProbeDetailsSidebarPanel.prototype.set inspectedProbeSets):
(WI.ProbeDetailsSidebarPanel.prototype.initialLayout):
Add checks to ensure that DOM elements for each probe section exist before trying to
add/remove them from the sidebar. This can happen if probes are inspected before the sidebar
is shown for the first time.

* UserInterface/Views/ProbeSetDataGrid.js:
(WI.ProbeSetDataGrid):
(WI.ProbeSetDataGrid.columnIdentifierForProbe): Added.
(WI.ProbeSetDataGrid.prototype._setupProbe):
(WI.ProbeSetDataGrid.prototype._teardownProbe):
(WI.ProbeSetDataGrid.prototype._probeExpressionChanged):
* UserInterface/Views/ProbeSetDataGridNode.js:
(WI.ProbeSetDataGridNode.prototype.set frame):
Provide better column identifiers for each probe's `WI.DataGrid`. It's possible for the
numeric probe ID value to be stringified when passing it into the constructor of
`WI.DataGrid`, which will not match the original numeric value on later retrieval.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234889 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoWebKitTestRunner should support watch devices
jbedard@apple.com [Wed, 15 Aug 2018 15:41:30 +0000 (15:41 +0000)]
WebKitTestRunner should support watch devices
https://bugs.webkit.org/show_bug.cgi?id=188570

Reviewed by Alex Christensen.

* WebKitTestRunner/Configurations/WebKitTestRunnerApp.xcconfig:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234888 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[WPE][GTK] WaylandCompositor fails to properly remove surface from its page map
mcatanzaro@igalia.com [Wed, 15 Aug 2018 15:17:55 +0000 (15:17 +0000)]
[WPE][GTK] WaylandCompositor fails to properly remove surface from its page map
https://bugs.webkit.org/show_bug.cgi?id=188520

Reviewed by Alex Christensen.

Source/WebKit:

willDestroySurface overwrites the surface pointer in the map's iterator in an attempt to
change the value of the surface pointer in the map, but it doesn't work because changing
the iterator does not change the map itself. There's no need to fix this function: it's
better to use WeakPtr instead.

* UIProcess/gtk/WaylandCompositor.cpp:
(WebKit::WaylandCompositor::getTexture):
(WebKit::WaylandCompositor::bindSurfaceToWebPage):
(WebKit::WaylandCompositor::unregisterWebPage):
(WebKit::WaylandCompositor::willDestroySurface): Deleted.
* UIProcess/gtk/WaylandCompositor.h:

Source/WTF:

Add a comment pointing to CanMakeWeakPtr, since it's useful and I very nearly missed it.

* wtf/WeakPtr.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234887 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoLayout Test editing/input/press-tab-during-ime-composition.html is failing
pvollan@apple.com [Wed, 15 Aug 2018 14:32:06 +0000 (14:32 +0000)]
Layout Test editing/input/press-tab-during-ime-composition.html is failing
https://bugs.webkit.org/show_bug.cgi?id=188600

Unreviewed test gardening.

* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234886 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[Attachment SPI] Remove attachment display mode options
wenson_hsieh@apple.com [Wed, 15 Aug 2018 14:26:05 +0000 (14:26 +0000)]
[Attachment SPI] Remove attachment display mode options
https://bugs.webkit.org/show_bug.cgi?id=188596

Reviewed by Dan Bernstein.

Source/WebCore:

Remove the ability to specify an "in-place" or "icon" representation for attachment elements, as well as logic
needed to allow an attachment element to render image or video elements in a shadow root. The requirements that
initially drove this effort are obviated by r227068, which allows Mail to intercept and provide a custom scheme
for images inserted into the document via rich editing operations.

Removed some existing API tests in _WKAttachmentTests that exercised this functionality.

* editing/Editor.cpp:
(WebCore::Editor::insertAttachmentFromFile):
* editing/cocoa/WebContentReaderCocoa.mm:
(WebCore::createFragmentForImageAttachment):
(WebCore::replaceRichContentWithAttachments):
* html/AttachmentTypes.h:
(WebCore::AttachmentDisplayOptions::encode const):
(WebCore::AttachmentDisplayOptions::decode):

Remove logic for encoding the attachment display mode. While this leaves AttachmentDisplayOptions completely
empty, I haven't removed AttachmentDisplayOptions as well in this patch, since it's not clear that we won't be
needing any mechanism for influencing the display of attachment elements inserted via native SPI.

(): Deleted.
* html/HTMLAttachmentElement.cpp:
(WebCore::HTMLAttachmentElement::createElementRenderer):
(WebCore::HTMLAttachmentElement::setFile):
(WebCore::HTMLAttachmentElement::parseAttribute):
(WebCore::HTMLAttachmentElement::invalidateShadowRootChildrenIfNecessary): Deleted.
(WebCore::HTMLAttachmentElement::attachmentRenderer const): Deleted.
(WebCore::HTMLAttachmentElement::updateDisplayMode): Deleted.
(WebCore::HTMLAttachmentElement::ensureInnerImage): Deleted.
(WebCore::HTMLAttachmentElement::ensureInnerVideo): Deleted.
(WebCore::HTMLAttachmentElement::innerImage const): Deleted.
(WebCore::HTMLAttachmentElement::innerVideo const): Deleted.
(WebCore::HTMLAttachmentElement::populateShadowRootIfNecessary): Deleted.

Remove logic for building the shadow root to house inline media elements.

* html/HTMLAttachmentElement.h:

The renderer of the attachment element is once again always a `RenderAttachment`, so we can remove
`attachmentRenderer()` altogether and revert to overriding `renderer()` to return a `RenderAttachment*`.

* page/DragController.cpp:
(WebCore::DragController::startDrag):
* rendering/RenderAttachment.h:
(WebCore::HTMLAttachmentElement::renderer const):

Source/WebKit:

Remove attachment display mode from WebKit. Note that _WKAttachmentDisplayOptions needs to remain in the private
header for source compatibility with Mail.

* UIProcess/API/Cocoa/_WKAttachment.mm:
(-[_WKAttachmentDisplayOptions coreDisplayOptions]): Deleted.
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::setAttachmentDisplayOptions):

Tools:

Remove API tests and API test helpers for verifying the behavior of in-place media attachment elements.

* TestWebKitAPI/Tests/WebKitCocoa/WKAttachmentTests.mm:
(-[TestWKWebView synchronouslyInsertAttachmentWithFilename:contentType:data:]):
(TestWebKitAPI::TEST):
(testVideoData): Deleted.
(displayOptionsWithMode): Deleted.
(-[TestWKWebView synchronouslyInsertAttachmentWithFilename:contentType:data:options:]): Deleted.

LayoutTests:

Remove a layout test which verified that attachment elements with `-webkit-appearance: none` would render
child elements. This was only used to implement in-place attachment display modes.

* fast/attachment/attachment-without-appearance-expected.html: Removed.
* fast/attachment/attachment-without-appearance.html: Removed.
* platform/gtk/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234885 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[IntersectionObserver] Do not hold a strong reference to the root element
ajuma@chromium.org [Wed, 15 Aug 2018 14:13:06 +0000 (14:13 +0000)]
[IntersectionObserver] Do not hold a strong reference to the root element
https://bugs.webkit.org/show_bug.cgi?id=188575

Reviewed by Simon Fraser.

Source/WebCore:

Make IntersectionObserver have only a raw pointer to its root element rather than
a reference, so that an otherwise-unreachable root isn't kept alive. Add logic to
to clear this pointer when the root element gets deleted.

Test: intersection-observer/root-element-deleted.html

* dom/Element.cpp:
(WebCore::Element::~Element):
(WebCore::Element::disconnectFromIntersectionObservers):
(WebCore::Element::ensureIntersectionObserverData):
(WebCore::Element::intersectionObserverData):
* dom/Element.h:
* dom/ElementRareData.cpp:
* dom/ElementRareData.h:
(WebCore::ElementRareData::intersectionObserverData):
(WebCore::ElementRareData::setIntersectionObserverData):
* page/IntersectionObserver.cpp:
(WebCore::IntersectionObserver::create):
(WebCore::IntersectionObserver::IntersectionObserver):
(WebCore::IntersectionObserver::~IntersectionObserver):
(WebCore::IntersectionObserver::rootDestroyed):
* page/IntersectionObserver.h:
(WebCore::IntersectionObserver::root const):

LayoutTests:

* intersection-observer/root-element-deleted-expected.txt: Added.
* intersection-observer/root-element-deleted.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234884 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[Nicosia] Add Nicosia::BackingStoreTextureMapperImpl
zandobersek@gmail.com [Wed, 15 Aug 2018 06:15:23 +0000 (06:15 +0000)]
[Nicosia] Add Nicosia::BackingStoreTextureMapperImpl
https://bugs.webkit.org/show_bug.cgi?id=188548

Reviewed by Carlos Garcia Campos.

Add the Nicosia::BackingStoreTextureMapperImpl class, the
TextureMapper-specific implementation that will extend the BackingStore
class.

Purpose of this class is to manage content of painted layers. In the
LayerState object that will be exposed to the CoordinatedGraphicsLayer
owner we keep the current and previous TiledBackingStore objects that
will be used for painting. A TileUpdate object is used to store all tile
creation, removal and update changes that happen during the painting
that's performed during the CoordinatedGraphicsLayer flush.

At the point of synchronization under the Nicosia::Scene object, these
updates will be moved over to the pending TileUpdate object that will
then be accessed during the composition step. For that purpose we keep
in the CompositionState member object a reference to the
CoordinatedBackingStore object that will get updated with the tile
creation, removal and update changes (if there are any). The composition
step will use the takeUpdate() method to retrieve all these changes and
apply them outside of the point of synchronization (avoiding this sync
step to take too long).

This will be integrated into the CoordinatedGraphicsLayer class at a
later point, when the switch to the new infrastructure can be made in
one go.

* platform/TextureMapper.cmake:
* platform/graphics/nicosia/texmap/NicosiaBackingStoreTextureMapperImpl.cpp: Added.
(Nicosia::BackingStoreTextureMapperImpl::createFactory):
(Nicosia::BackingStoreTextureMapperImpl::tiledBackingStoreHasPendingTileCreation):
(Nicosia::BackingStoreTextureMapperImpl::createTile):
(Nicosia::BackingStoreTextureMapperImpl::updateTile):
(Nicosia::BackingStoreTextureMapperImpl::removeTile):
(Nicosia::BackingStoreTextureMapperImpl::flushUpdate):
(Nicosia::BackingStoreTextureMapperImpl::takeUpdate):
* platform/graphics/nicosia/texmap/NicosiaBackingStoreTextureMapperImpl.h: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234883 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoWeb Inspector: Table should not center rows when scrolling them into view
mattbaker@apple.com [Wed, 15 Aug 2018 06:10:34 +0000 (06:10 +0000)]
Web Inspector: Table should not center rows when scrolling them into view
https://bugs.webkit.org/show_bug.cgi?id=188593
<rdar://problem/43311660>

Reviewed by Devin Rousso.

* UserInterface/Views/Table.js:
(WI.Table.prototype._handleKeyDown):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234882 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoUnreviewed, change my emails.
drousso@apple.com [Wed, 15 Aug 2018 04:47:16 +0000 (04:47 +0000)]
Unreviewed, change my emails.

* Scripts/webkitpy/common/config/contributors.json:
* Scripts/webkitpy/common/config/watchlist:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234881 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoUnreviewed, rolling out r234874 and r234876.
Hironori.Fujii@sony.com [Wed, 15 Aug 2018 04:46:12 +0000 (04:46 +0000)]
Unreviewed, rolling out r234874 and r234876.

WinCairo port can't compile

Reverted changesets:

"[JSC] Add GPRReg::InvalidGPRReg and FPRReg::InvalidFPRReg"
https://bugs.webkit.org/show_bug.cgi?id=188589
https://trac.webkit.org/changeset/234874

"Unreviewed, attempt to fix CLoop build"
https://bugs.webkit.org/show_bug.cgi?id=188589
https://trac.webkit.org/changeset/234876

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234880 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoHashMap<Ref<P>, V> asserts when V is not zero for its empty value
sbarati@apple.com [Wed, 15 Aug 2018 04:08:08 +0000 (04:08 +0000)]
HashMap<Ref<P>, V> asserts when V is not zero for its empty value
https://bugs.webkit.org/show_bug.cgi?id=188582

Reviewed by Sam Weinig.

Source/JavaScriptCore:

* runtime/SparseArrayValueMap.h:

Source/WTF:

The issue happened when we'd fill the hash table buffer with empty values. We
would iterate the buffer and invoke placement new with the incoming value being the
empty value. For Ref, this means that, we'd call its move constructor, which calls
leakRef(), which asserts that the Ref's pointer is not null. We'd like to keep
this assert since it catches bugs where you leakRef() more than once or WTFMove
an already moved Ref.

This patch fixes this issue by adding a new trait for constructing an empty
value. We use that in HashTable instead of directly calling placement new.

* wtf/HashTable.h:
(WTF::HashTableBucketInitializer<false>::initialize):
* wtf/HashTraits.h:
(WTF::GenericHashTraits::constructEmptyValue):
(WTF::HashTraits<Ref<P>>::constructEmptyValue):
(WTF::KeyValuePairHashTraits::constructEmptyValue):

Tools:

* TestWebKitAPI/Tests/WTF/HashMap.cpp:
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234879 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[LFC][Floating] Add support for negative clearance.
zalan@apple.com [Wed, 15 Aug 2018 02:59:51 +0000 (02:59 +0000)]
[LFC][Floating] Add support for negative clearance.
https://bugs.webkit.org/show_bug.cgi?id=188555

Reviewed by Simon Fraser.

Source/WebCore:

1. Compute clearance to avoid float(s) (border box needs to avoid floats)
2. Reset vertical margins to non-collapsed values.
4. Adjust clearance with the new margins.
5. Take the adjusted clearance and move the box vertically if needed.

Test: fast/block/block-only/margin-collapse-with-clearance.html

* layout/FloatingContext.cpp:
(WebCore::Layout::FloatingContext::verticalPositionWithClearance const):
* layout/FloatingState.cpp:
(WebCore::Layout::FloatingState::FloatItem::FloatItem):
* layout/FormattingContext.cpp:
(WebCore::Layout::FormattingContext::mapBoxToAncestor):
(WebCore::Layout::FormattingContext::mapTopLeftToAncestor):
(WebCore::Layout::FormattingContext::mapCoordinateToAncestor):
(WebCore::Layout::FormattingContext::mapToAncestor): Deleted.
* layout/FormattingContext.h:
* layout/LayoutUnits.h:
(WebCore::Layout::Position::Position):
(WebCore::Layout::Position::moveBy):
* layout/displaytree/DisplayBox.h:

Tools:

* LayoutReloaded/misc/LFC-passing-tests.txt:

LayoutTests:

* fast/block/block-only/margin-collapse-with-clearance-expected.txt: Added.
* fast/block/block-only/margin-collapse-with-clearance.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234878 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoUnreviewed, rolling out r234859.
Hironori.Fujii@sony.com [Wed, 15 Aug 2018 02:42:59 +0000 (02:42 +0000)]
Unreviewed, rolling out r234859.

Windows ports can't compile

Reverted changeset:

"Use a Variant instead of a union in CSSSelector"
https://bugs.webkit.org/show_bug.cgi?id=188559
https://trac.webkit.org/changeset/234859

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234877 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoUnreviewed, attempt to fix CLoop build
yusukesuzuki@slowstart.org [Wed, 15 Aug 2018 01:18:17 +0000 (01:18 +0000)]
Unreviewed, attempt to fix CLoop build
https://bugs.webkit.org/show_bug.cgi?id=188589

* assembler/MacroAssembler.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234876 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoCrash in WebKit::filterPreloadHSTSEntry via NetworkProcess::getHostNamesWithHSTSCache
sihui_liu@apple.com [Wed, 15 Aug 2018 00:55:19 +0000 (00:55 +0000)]
Crash in WebKit::filterPreloadHSTSEntry via NetworkProcess::getHostNamesWithHSTSCache
https://bugs.webkit.org/show_bug.cgi?id=188576
<rdar://problem/43148977>

Reviewed by Alex Christensen.

Source/WebKit:

* NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::fetchWebsiteData):

Tools:

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKitCocoa/WKWebsiteDatastore.mm: Added.
(TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234875 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[JSC] Add GPRReg::InvalidGPRReg and FPRReg::InvalidFPRReg
yusukesuzuki@slowstart.org [Wed, 15 Aug 2018 00:50:50 +0000 (00:50 +0000)]
[JSC] Add GPRReg::InvalidGPRReg and FPRReg::InvalidFPRReg
https://bugs.webkit.org/show_bug.cgi?id=188589

Reviewed by Mark Lam.

Source/JavaScriptCore:

Since GPRReg(RegisterID) and FPRReg(FPRegisterID) do not include -1 in their enum values,
UBSan dumps bunch of warnings "runtime error: load of value 4294967295, which is not a valid value for type 'RegisterID'".

1. We add InvalidGPRReg and InvalidFPRReg to enum values of GPRReg and FPRReg to suppress the above warnings.
2. We make GPRReg and FPRReg int8_t enums.
3. We replace `#define InvalidGPRReg ((JSC::GPRReg)-1)` to `static constexpr GPRReg InvalidGPRReg { GPRReg::InvalidGPRReg };`.

* assembler/ARM64Assembler.h:
* assembler/ARMAssembler.h:
* assembler/ARMv7Assembler.h:
* assembler/MIPSAssembler.h:
* assembler/X86Assembler.h:
* jit/FPRInfo.h:
* jit/GPRInfo.h:
(JSC::JSValueRegs::JSValueRegs):
(JSC::JSValueRegs::tagGPR const):
(JSC::JSValueRegs::payloadGPR const):
(JSC::JSValueSource::JSValueSource):
(JSC::JSValueSource::unboxedCell):
(JSC::JSValueSource::operator bool const):
(JSC::JSValueSource::base const):
(JSC::JSValueSource::tagGPR const):
(JSC::JSValueSource::payloadGPR const):
(JSC::JSValueSource::hasKnownTag const):

Source/WebCore:

No behavior change.

* cssjit/FunctionCall.h:
(WebCore::FunctionCall::FunctionCall):
* cssjit/SelectorCompiler.cpp:
(WebCore::SelectorCompiler::SelectorCodeGenerator::modulo):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234874 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoisValidCSSSelector is unsafe to be called from a non-main thread
achristensen@apple.com [Wed, 15 Aug 2018 00:30:50 +0000 (00:30 +0000)]
isValidCSSSelector is unsafe to be called from a non-main thread
https://bugs.webkit.org/show_bug.cgi?id=188581
<rdar://problem/40517358>

Reviewed by Sam Weinig.

Source/WebCore:

Parsing and determining whether the css selectors are valid is fast enough to do before
hopping to the background thread for the slow NFA/DFA operations and writing to disk.
Doing it on the main thread avoids the thread safety issues in the CSSParser's use of strings.

* contentextensions/ContentExtensionCompiler.cpp:
(WebCore::ContentExtensions::compileRuleList):
* contentextensions/ContentExtensionCompiler.h:
* contentextensions/ContentExtensionParser.cpp:
(WebCore::ContentExtensions::isValidCSSSelector):
(WebCore::ContentExtensions::loadEncodedRules):
(WebCore::ContentExtensions::parseRuleList):
* contentextensions/ContentExtensionParser.h:
* contentextensions/ContentExtensionRule.cpp:
(WebCore::ContentExtensions::Trigger::isolatedCopy const):
(WebCore::ContentExtensions::Action::isolatedCopy const):
* contentextensions/ContentExtensionRule.h:
(WebCore::ContentExtensions::Trigger::isEmpty const):
(WebCore::ContentExtensions::Trigger::operator== const):
(WebCore::ContentExtensions::Action::Action):
(WebCore::ContentExtensions::ContentExtensionRule::isolatedCopy const):
(WebCore::ContentExtensions::ContentExtensionRule::operator== const):
(WebCore::ContentExtensions::vectorIsolatedCopy):

Source/WebKit:

* UIProcess/API/APIContentRuleListStore.cpp:
(API::compiledToFile):
(API::ContentRuleListStore::lookupContentRuleList):
(API::ContentRuleListStore::getAvailableContentRuleListIdentifiers):
(API::ContentRuleListStore::compileContentRuleList):
(API::ContentRuleListStore::removeContentRuleList):
(API::ContentRuleListStore::getContentRuleListSource):
* UIProcess/API/APIContentRuleListStore.h:
* UIProcess/API/Cocoa/WKContentRuleListStore.mm:

Source/WTF:

* wtf/Vector.h:
(WTF::minCapacity>::isolatedCopy):

Tools:

* TestWebKitAPI/Tests/WebCore/ContentExtensions.cpp:
(TestWebKitAPI::InMemoryCompiledContentExtension::create):
(TestWebKitAPI::checkCompilerError):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234873 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoNSURLAuthenticationMethodOAuth challenges are surfaced to clients in -didReceiveAuthe...
commit-queue@webkit.org [Wed, 15 Aug 2018 00:10:12 +0000 (00:10 +0000)]
NSURLAuthenticationMethodOAuth challenges are surfaced to clients in -didReceiveAuthenticationChallenge as NSURLAuthenticationMethodDefault
https://bugs.webkit.org/show_bug.cgi?id=186870
Source/WebCore:

<rdar://problem/41314410>

Patch by Ansh Shukla <ansh_shukla@apple.com> on 2018-08-14
Reviewed by Alex Christensen.

Add the ProtectionSpaceAuthenticationSchemeOAuth type.

* platform/network/ProtectionSpaceBase.cpp:
(WebCore::ProtectionSpaceBase::isPasswordBased const): Return yes because the oauth challenge
expects a token in return.
* platform/network/ProtectionSpaceBase.h:
* platform/network/cocoa/ProtectionSpaceCocoa.mm:
(WebCore::scheme):
(WebCore::ProtectionSpace::nsSpace const):

Source/WebCore/PAL:

<rdar://problem/41314410>

Patch by Ansh Shukla <ansh_shukla@apple.com> on 2018-08-14
Reviewed by Alex Christensen.

* pal/spi/cf/CFNetworkSPI.h: Declare OAuth string when not building against the
internal SDK.

Source/WebKit:

<rdar://problem/41314410>

Patch by Ansh Shukla <ansh_shukla@apple.com> on 2018-08-14
Reviewed by Alex Christensen.

Correctly expose the OAuth protection space type in API.

* UIProcess/API/C/WKAPICast.h:
(WebKit::toAPI):
* UIProcess/API/C/WKProtectionSpaceTypes.h:

Tools:

<rdar://problem/41314410>

Patch by Ansh Shukla <ansh_shukla@apple.com> on 2018-08-14
Reviewed by Alex Christensen.

* WebKitTestRunner/TestController.cpp:
(WTR::toString):
(WTR::TestController::canAuthenticateAgainstProtectionSpace): Expose type of authentication challenge so we can test OAuth.
(WTR::TestController::didReceiveAuthenticationChallenge):

LayoutTests:

Patch by Ansh Shukla <ansh_shukla@apple.com> on 2018-08-14
Reviewed by Alex Christensen.

Ensure the exposed authentication type to clients is OAuth.

* http/tests/loading/oauth-expected.txt: Added.
* http/tests/loading/oauth.html: Added.
* http/tests/loading/resources/oauth-subresource.php: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234870 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoUnreviewed test gardening for mac-wk1.
ryanhaddad@apple.com [Tue, 14 Aug 2018 23:56:33 +0000 (23:56 +0000)]
Unreviewed test gardening for mac-wk1.

* platform/mac-wk1/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234869 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoMark legacy-animation-engine/animations/combo-transform-translate+scale.html as flaky.
ryanhaddad@apple.com [Tue, 14 Aug 2018 23:56:30 +0000 (23:56 +0000)]
Mark legacy-animation-engine/animations/combo-transform-translate+scale.html as flaky.
https://bugs.webkit.org/show_bug.cgi?id=185308

Unreviewed test gardening.

* TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234868 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[Curl] Test gardening
Basuke.Suzuki@sony.com [Tue, 14 Aug 2018 23:34:49 +0000 (23:34 +0000)]
[Curl] Test gardening
https://bugs.webkit.org/show_bug.cgi?id=188587

Unreviewed test gardening.

Added platform dependent expected result.

* platform/wincairo/http/tests/xmlhttprequest/methods-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234867 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoAdd MotionMark 1.1 to browserbench
rniwa@webkit.org [Tue, 14 Aug 2018 23:26:06 +0000 (23:26 +0000)]
Add MotionMark 1.1 to browserbench
https://bugs.webkit.org/show_bug.cgi?id=188585
<rdar://problem/43308076>

Patch by Jon Lee <jonlee@apple.com> on 2018-08-14
Rubber-stamped by Ryosuke Niwa.

* MotionMark1.1/: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234866 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoRemove api misuse check so that custom webcontent service identifier can be set at...
commit-queue@webkit.org [Tue, 14 Aug 2018 23:22:09 +0000 (23:22 +0000)]
Remove api misuse check so that custom webcontent service identifier can be set at runtime
https://bugs.webkit.org/show_bug.cgi?id=188579

Patch by Ben Richards <benton_richards@apple.com> on 2018-08-14
Reviewed by Ryosuke Niwa.

Changed API misuse check so that a custom bundle identifier can be set at runtime with a debug flag

* UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::setCustomWebContentServiceBundleIdentifier):
* UIProcess/WebProcessProxy.cpp:
(WebKit::WebProcessProxy::getLaunchOptions):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234865 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[Curl] Implement platform default timeout interval.
Basuke.Suzuki@sony.com [Tue, 14 Aug 2018 23:17:13 +0000 (23:17 +0000)]
[Curl] Implement platform default timeout interval.
https://bugs.webkit.org/show_bug.cgi?id=188565

Reviewed by Alex Christensen.

Source/WebCore:

Curl port didn't implement platform default timeout interval. It treated
zero value for timeout interval as no timeout.

Add platform dependent timeout interval into CurlContext and use that if
timeout and default timeout aren't supplied.

Tests: http/tests/xmlhttprequest/on-network-timeout-error-during-preflight.html

* platform/network/curl/CurlContext.cpp:
(WebCore::CurlHandle::setTimeout):
* platform/network/curl/CurlContext.h:
(WebCore::CurlContext::defaultTimeoutInterval const):
* platform/network/curl/CurlRequest.cpp:
(WebCore::CurlRequest::setupTransfer):
(WebCore::CurlRequest::timeoutInterval const):
(WebCore::CurlRequest::didCompleteTransfer):
* platform/network/curl/CurlRequest.h:

LayoutTests:

* platform/wincairo/TestExpectations:
  - xmlhttprequest/on-network-timeout-error-during-preflight.html

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234864 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[webkitpy][Win] LayoutTests: test names should be Unix style, separated by slash...
ross.kirsling@sony.com [Tue, 14 Aug 2018 22:43:39 +0000 (22:43 +0000)]
[webkitpy][Win] LayoutTests: test names should be Unix style, separated by slash not backslash
https://bugs.webkit.org/show_bug.cgi?id=187973

Patch by Fujii Hironori <Hironori.Fujii@sony.com> on 2018-08-14
Reviewed by Alex Christensen.

In LayoutTests, test names have been canonicalized in Unix style
since Bug 63597, for example 'fast/css/001.html'. But Bug 179219,
Bug 179572, Bug 180660, and Bug 181814 have changed to use
os.path.seq instead of slash if Windows Python is used.

Revert parts of those changes. Change relative_test_filename to
return a slash-separated test name as chromium_win.py used to do.

This change fixes all 41 test-webkitpy failures in WinCairo port.

* Scripts/webkitpy/layout_tests/models/test_expectations.py:
(TestExpectationParser._parse_line): Do not convert test names with normpath.
* Scripts/webkitpy/port/base.py:
(Port.normalize_test_name): Use TEST_PATH_SEPARATOR instead of os.path.sep.
(Port.relative_test_filename): Replace self.host.filesystem.sep with self.TEST_PATH_SEPARATOR.
(Port.abspath_for_test): Replace self.TEST_PATH_SEPARATOR with self.host.filesystem.sep.
* Scripts/webkitpy/port/driver.py:
(Driver): Use '/' instead of os.sep.
* Scripts/webkitpy/port/win.py:
(WinCairoPort): Do not override TEST_PATH_SEPARATOR.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234863 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months agoMark svg/custom/scrolling-embedded-svg-file-image-repaint-problem.html as flaky.
ryanhaddad@apple.com [Tue, 14 Aug 2018 21:15:16 +0000 (21:15 +0000)]
Mark svg/custom/scrolling-embedded-svg-file-image-repaint-problem.html as flaky.
https://bugs.webkit.org/show_bug.cgi?id=188578

Unreviewed test gardening.

* platform/ios/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234862 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[ews-build] Add build step to run WK1 layout-test
aakash_jain@apple.com [Tue, 14 Aug 2018 20:13:39 +0000 (20:13 +0000)]
[ews-build] Add build step to run WK1 layout-test
https://bugs.webkit.org/show_bug.cgi?id=188498

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-build/steps.py:
(RunWebKit1Tests): Class to run WebKit1Tests.
* BuildSlaveSupport/ews-build/steps_unittest.py: Added unit-tests.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234861 268f45cc-cd09-0410-ab3c-d52691b4dbfc

15 months ago[Flatpak] Fix --cmakeargs
commit-queue@webkit.org [Tue, 14 Aug 2018 19:04:56 +0000 (19:04 +0000)]
[Flatpak] Fix --cmakeargs
https://bugs.webkit.org/show_bug.cgi?id=188567

Patch by Patrick Griffis <pgriffis@igalia.com> on 2018-08-14
Reviewed by Michael Catanzaro.

* flatpak/flatpakutils.py:
(WebkitFlatpak.setup_dev_env):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234860 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 months agoUse a Variant instead of a union in CSSSelector
achristensen@apple.com [Tue, 14 Aug 2018 18:24:02 +0000 (18:24 +0000)]
Use a Variant instead of a union in CSSSelector
https://bugs.webkit.org/show_bug.cgi?id=188559

Reviewed by Antti Koivisto.

Source/WebCore:

No change in behavior. This just makes some of the existing problems more obvious and easy to fix.

I moved m_caseInsensitiveAttributeValueMatching to RareData because it's only used with RareData.
I only have m_isForPage when assertions are enabled because it's only used for an assertion.
The rest is pretty straightforward translating union syntax to Variant syntax.
I use RefPtr for now where I could use Ref because it's never null to make copying easier, but that's temporary.

* css/CSSSelector.cpp:
(WebCore::CSSSelector::CSSSelector):
(WebCore::CSSSelector::createRareData):
(WebCore::CSSSelector::setAttribute):
(WebCore::CSSSelector::setArgument):
(WebCore::CSSSelector::setLangArgumentList):
(WebCore::CSSSelector::setSelectorList):
(WebCore::CSSSelector::setNth):
(WebCore::CSSSelector::matchNth const):
(WebCore::CSSSelector::nthA const):
(WebCore::CSSSelector::nthB const):
(WebCore::CSSSelector::RareData::RareData):
* css/CSSSelector.h:
(WebCore::CSSSelector::argument const):
(WebCore::CSSSelector::langArgumentList const):
(WebCore::CSSSelector::selectorList const):
(WebCore::CSSSelector::attribute const):
(WebCore::CSSSelector::attributeCanonicalLocalName const):
(WebCore::CSSSelector::setValue):
(WebCore::CSSSelector::CSSSelector):
(WebCore::CSSSelector::~CSSSelector):
(WebCore::CSSSelector::tagQName const):
(WebCore::CSSSelector::tagLowercaseLocalName const):
(WebCore::CSSSelector::value const):
(WebCore::CSSSelector::serializingValue const):
(WebCore::CSSSelector::attributeValueMatchingIsCaseInsensitive const):
(WebCore::CSSSelector::RareData::create): Deleted.
* css/parser/CSSParserImpl.cpp:
(WebCore::CSSParserImpl::parsePageSelector):
* css/parser/CSSParserSelector.h:

Source/WTF:

* wtf/Variant.h:
Add packing macros to make it so Variant-containing structures don't always have 7 bytes of padding per Variant.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234859 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 months agoAdd missing availability macro.
keith_miller@apple.com [Tue, 14 Aug 2018 17:59:32 +0000 (17:59 +0000)]
Add missing availability macro.
https://bugs.webkit.org/show_bug.cgi?id=188563

Reviewed by Mark Lam.

* API/JSValueRef.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234858 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 months agoUnhandled Promise Rejection logging in workers should not emit ErrorEvent to host...
yusukesuzuki@slowstart.org [Tue, 14 Aug 2018 17:56:27 +0000 (17:56 +0000)]
Unhandled Promise Rejection logging in workers should not emit ErrorEvent to host Worker object
https://bugs.webkit.org/show_bug.cgi?id=188551

Reviewed by Youenn Fablet.

LayoutTests/imported/w3c:

* web-platform-tests/WebCryptoAPI/derive_bits_keys/ecdh_bits.https.worker-expected.txt:
* web-platform-tests/WebCryptoAPI/derive_bits_keys/ecdh_keys.https.worker-expected.txt:
* web-platform-tests/streams/readable-streams/tee.dedicatedworker-expected.txt:

Source/WebCore:

Previously we dispatched ErrorEvent on the Worker object of the host side when the unhandled promise
rejection happens in the worker. But that was wrong. We should not dispatch such an event and we
should just log the error message.

* dom/ScriptExecutionContext.cpp:
(WebCore::ScriptExecutionContext::reportUnhandledPromiseRejection):

LayoutTests:

* http/wpt/workers/promise-unhandled-rejection.any.worker-expected.txt:
* js/dom/unhandled-promise-rejection-bindings-type-error-in-workers-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234857 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 months agoEnhancement request: Make export-w3c-test-changes add the PR to "See Also" links
commit-queue@webkit.org [Tue, 14 Aug 2018 17:54:56 +0000 (17:54 +0000)]
Enhancement request: Make export-w3c-test-changes add the PR to "See Also" links
https://bugs.webkit.org/show_bug.cgi?id=186140

Patch by Darshan Kadu <dkadu@igalia.com> on 2018-08-14
Reviewed by Youenn Fablet.

Added an optional parameter see_also to post_comment_to_bug function and used it for adding see_also variable's content in "See Also"
* Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
(Bugzilla.post_comment_to_bug):
Modified mock for see_also parameter
* Scripts/webkitpy/common/net/bugzilla/bugzilla_mock.py:
(MockBugzilla.post_comment_to_bug):
Made following unit tests to have see_also in bug comment
* Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
* Scripts/webkitpy/tool/bot/sheriff_unittest.py:
(SheriffTest.test_post_blame_comment_on_bug):
* Scripts/webkitpy/tool/commands/applywatchlistlocal_unittest.py:
* Scripts/webkitpy/tool/commands/queues_unittest.py:
* Scripts/webkitpy/tool/commands/upload_unittest.py:
* Scripts/webkitpy/tool/steps/applywatchlist_unittest.py:
* Scripts/webkitpy/w3c/test_exporter.py:
(WebPlatformTestExporter.make_pull_request):
* Scripts/webkitpy/w3c/test_exporter_unittest.py:
(TestExporterTest.MockBugzilla.post_comment_to_bug):
(TestExporterTest.test_export):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234856 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 months ago[JSC] GetByIdStatus::m_wasSeenInJIT is touched in GetByIdStatus::slowVersion
yusukesuzuki@slowstart.org [Tue, 14 Aug 2018 17:46:35 +0000 (17:46 +0000)]
[JSC] GetByIdStatus::m_wasSeenInJIT is touched in GetByIdStatus::slowVersion
https://bugs.webkit.org/show_bug.cgi?id=188560

Reviewed by Keith Miller.

While GetByIdStatus() / GetByIdStatus(status) constructors do not set m_wasSeenInJIT,
it is loaded unconditionally in GetByIdStatus::slowVersion. This access to the
uninitialized member field is caught in UBSan. This patch fixes it by adding an initializer
`m_wasSeenInJIT { false }`.

* bytecode/GetByIdStatus.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234855 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 months agoUnreviewed, follow-up patch for comments
yusukesuzuki@slowstart.org [Tue, 14 Aug 2018 17:45:09 +0000 (17:45 +0000)]
Unreviewed, follow-up patch for comments
https://bugs.webkit.org/show_bug.cgi?id=188265

* bindings/js/JSExecState.cpp:
(WebCore::JSExecState::didLeaveScriptContext):
Do nothing if `context` is nullptr. It is OK since we do not need to drain microtasks / rejected
promise events after ScriptExecutionContext is gone.

* dom/Microtasks.cpp:
(WebCore::MicrotaskQueue::contextQueue):
Drop unnecessary assertion since it is subsumed by downcast<>.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234854 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 months ago[DFG] DFGPredictionPropagation should set PrimaryPass when processing invariants
yusukesuzuki@slowstart.org [Tue, 14 Aug 2018 17:23:50 +0000 (17:23 +0000)]
[DFG] DFGPredictionPropagation should set PrimaryPass when processing invariants
https://bugs.webkit.org/show_bug.cgi?id=188557

Reviewed by Mark Lam.

DFGPredictionPropagationPhase should set PrimaryPass before processing invariants since
processing for ArithRound etc.'s invariants requires `m_pass` load. This issue is found
in UBSan's result.

* dfg/DFGPredictionPropagationPhase.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234853 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 months ago[JSC] Should not rotate constant with 64
yusukesuzuki@slowstart.org [Tue, 14 Aug 2018 17:16:31 +0000 (17:16 +0000)]
[JSC] Should not rotate constant with 64
https://bugs.webkit.org/show_bug.cgi?id=188556

Reviewed by Mark Lam.

To defend against JIT splaying, we rotate a constant with a randomly generated seed.
But if a seed becomes 64, the following code performs `value << 64` where value's type
is uint64_t, and it causes undefined behaviors (UBs). This patch limits the seed in the
range of [0, 64) not to generate code causing UBs. This is found by UBSan.

* assembler/MacroAssembler.h:
(JSC::MacroAssembler::generateRotationSeed):
(JSC::MacroAssembler::rotationBlindConstant):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234852 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 months agoFollow-up: [IntersectionObserver] Implement rootMargin parsing
ajuma@chromium.org [Tue, 14 Aug 2018 17:12:14 +0000 (17:12 +0000)]
Follow-up: [IntersectionObserver] Implement rootMargin parsing
https://bugs.webkit.org/show_bug.cgi?id=188469

Address review feedback from Darin Adler.

* page/IntersectionObserver.cpp:
(WebCore::IntersectionObserver::rootMargin const):
Use StringBuilder::appendLiteral instead of ::append to append a literal.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234851 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 months ago[ews-build] Add support for max_builds parameter for workers
aakash_jain@apple.com [Tue, 14 Aug 2018 16:53:08 +0000 (16:53 +0000)]
[ews-build] Add support for max_builds parameter for workers
https://bugs.webkit.org/show_bug.cgi?id=188531

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-build/loadConfig.py:
(loadBuilderConfig): Added support for max_builds.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234850 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 months agoRemoteLayerTreeTransaction should use OptionSet for change flags
antti@apple.com [Tue, 14 Aug 2018 16:48:45 +0000 (16:48 +0000)]
RemoteLayerTreeTransaction should use OptionSet for change flags
https://bugs.webkit.org/show_bug.cgi?id=188547

Reviewed by Simon Fraser.

* Shared/RemoteLayerTree/RemoteLayerTreePropertyApplier.mm:
(WebKit::RemoteLayerTreePropertyApplier::applyProperties):
* Shared/RemoteLayerTree/RemoteLayerTreeTransaction.h:
(WebKit::RemoteLayerTreeTransaction::LayerProperties::notePropertiesChanged):
(WebKit::RemoteLayerTreeTransaction::LayerProperties::resetChangedProperties):

Also remove unused everChangedProperties.

* Shared/RemoteLayerTree/RemoteLayerTreeTransaction.mm:
(WebKit::RemoteLayerTreeTransaction::LayerProperties::LayerProperties):
(WebKit::RemoteLayerTreeTransaction::LayerProperties::encode const):
(WebKit::RemoteLayerTreeTransaction::LayerProperties::decode):
* WebProcess/WebPage/RemoteLayerTree/PlatformCALayerRemote.cpp:
(WebKit::PlatformCALayerRemote::recursiveBuildTransaction):
* WebProcess/WebPage/RemoteLayerTree/RemoteLayerTreeDrawingArea.mm:
(WebKit::RemoteLayerTreeDrawingArea::flushLayers):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234849 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 months ago[Web Animations] Crash under AnimationTimeline::cancelOrRemoveDeclarativeAnimation()
graouts@webkit.org [Tue, 14 Aug 2018 15:15:07 +0000 (15:15 +0000)]
[Web Animations] Crash under AnimationTimeline::cancelOrRemoveDeclarativeAnimation()
https://bugs.webkit.org/show_bug.cgi?id=188519
<rdar://problem/43237889>

Reviewed by Eric Carlson.

Source/WebCore:

Test: webanimations/css-animation-effect-target-change-and-animation-removal-crash.html

We would crash because we blindly assumed an animation that was found in the previous style must be in the list of running animations
but in fact it could have been removed already due to the element being removed from the DOM or its effect target changing, etc. So when
we iterate over names of animations that were found in the previous style but not in the new style, we must make a null check to ensure
that there is an animation to remove. Adding an ASSERT() in AnimationTimeline::cancelOrRemoveDeclarativeAnimation() will also clarify the
expectation here.

* animation/AnimationTimeline.cpp:
(WebCore::AnimationTimeline::updateCSSAnimationsForElement):
(WebCore::AnimationTimeline::cancelOrRemoveDeclarativeAnimation):

LayoutTests:

Add a test where we clone the effect to be mutable and set a new target. At this stage the animation is no longer listed in the
m_elementToCSSAnimationByName map on AnimationTimeline. Then we remove the animation and force a style recalc for this element,
"anim" will be in the old style but not in the new style and we used to attempt to get an animation matching that name from
m_elementToCSSAnimationByName but it would be null, which would lead to a crash. Now we check that we indeed have such an animation
before proceeding.

* webanimations/css-animation-effect-target-change-and-animation-removal-crash-expected.html: Added.
* webanimations/css-animation-effect-target-change-and-animation-removal-crash.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234848 268f45cc-cd09-0410-ab3c-d52691b4dbfc

16 months agoSource/WebCore:
zalan@apple.com [Tue, 14 Aug 2018 14:22:45 +0000 (14:22 +0000)]
Source/WebCore:
[LFC][Floating] Adjust vertical position with non-collapsed previous sibling margin.
https://bugs.webkit.org/show_bug.cgi?id=188543

Reviewed by Antti Koivisto.

This patch ensures that the inital vertical position for a float is adjusted with the non-collapsed sibling margin.

<div id=A style="margin-bottom: 20px;"></div>
<div id=B style='float: left'></div>
<div id=C style="margin-top: 10px;"></div>

While computing the static position for element "B", we simply call marginBottom() on A.
In the case above, A's margin bottom is collapsed with C's margin top and the value is 0 (C.marginTop() is 20px).
However CSS spec says that in block formatting context, the non-collapsed margin should be used instead to offset the float box.
(The reason why this should not be part of the BlockMarginCollapse::marginBottom() logic is because it can not differentiate the context of
sibling float/sibling inflow. When we margin collapse, we always do it in the context of inflow boxes.)

Test: fast/block/block-only/float-and-siblings-with-margins.html

* layout/blockformatting/BlockFormattingContext.cpp:
(WebCore::Layout::BlockFormattingContext::layoutFormattingContextRoot const):
(WebCore::Layout::BlockFormattingContext::computeFloatingPosition const):
* layout/blockformatting/BlockFormattingContext.h:

Tools:
[LFC][Floating] Adjust vertical position with non-collapsing previous sibling margin.
https://bugs.webkit.org/show_bug.cgi?id=188543

Reviewed by Antti Koivisto.

* LayoutReloaded/misc/LFC-passing-tests.txt:

LayoutTests:
[LFC][Floating] Adjust vertical position with non-collapsing previous sibling margin.
https://bugs.webkit.org/show_bug.cgi?id=188543

Reviewed by Antti Koivisto.

* fast/block/block-only/float-and-siblings-with-margins-expected.txt: Added.
* fast/block/block-only/float-and-siblings-with-margins.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@234847 268f45cc-cd09-0410-ab3c-d52691b4dbfc