WebKit-https.git
4 years agoUnreviewed, rolling out r182247.
commit-queue@webkit.org [Thu, 2 Apr 2015 00:22:54 +0000 (00:22 +0000)]
Unreviewed, rolling out r182247.
https://bugs.webkit.org/show_bug.cgi?id=143326

unfortunately broke some tests (Requested by dino on #webkit).

Reverted changeset:

"Async loading of image resources"
https://bugs.webkit.org/show_bug.cgi?id=134488
http://trac.webkit.org/changeset/182247

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182266 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Content Extensions] Only add unique universal action locations.
achristensen@apple.com [Thu, 2 Apr 2015 00:16:36 +0000 (00:16 +0000)]
[Content Extensions] Only add unique universal action locations.
https://bugs.webkit.org/show_bug.cgi?id=143325

Reviewed by Benjamin Poulain.

* contentextensions/ContentExtensionCompiler.cpp:
(WebCore::ContentExtensions::compileRuleList):
Use a HashSet instead of a Vector for the universal action locations to remove duplicates.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182265 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agofast/events/mouse-cursor-image-set.html is timing out on Apple platforms
dino@apple.com [Wed, 1 Apr 2015 23:59:09 +0000 (23:59 +0000)]
fast/events/mouse-cursor-image-set.html is timing out on Apple platforms
https://bugs.webkit.org/show_bug.cgi?id=143323

Removing old flakiness setting now that we should be skipping.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182264 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFixes to material of inline media controls on OSX.
roger_fong@apple.com [Wed, 1 Apr 2015 23:47:35 +0000 (23:47 +0000)]
Fixes to material of inline media controls on OSX.
https://bugs.webkit.org/show_bug.cgi?id=143317
<rdar://problem/20357847>

Reviewed by Dean Jackson.

There are many existing tests for the media controls appearance.
They are currently skipped while we are still updating the controls.

Expand backgrounds of control panel and volume box.
Clip the darker edges away.
* Modules/mediacontrols/mediaControlsApple.css:
(audio::-webkit-media-controls-panel):
(audio::-webkit-media-controls-panel-background-container):
(audio::-webkit-media-controls-panel-tint):
(audio::-webkit-media-controls-panel-background):
(audio::-webkit-media-controls-panel .volume-box):
(audio::-webkit-media-controls-volume-slider-container-background):
(audio::-webkit-media-controls-volume-slider-container-tint):
(audio::-webkit-media-controls-panel .volume-box:active):
(video::-webkit-media-controls-volume-slider):
This needs to be above the background so it can be interacted with.
(video:-webkit-full-screen::-webkit-media-controls-panel .volume-box):
* Modules/mediacontrols/mediaControlsApple.js:
(Controller.prototype.createControls): Create a new div for the control material.
We need to clip the material and not the controls themselves.
(Controller.prototype.configureInlineControls): Use new div.
Remove some unnecessary logic involving hiding materials here.
It just causes the materials to turn on and off at the wrong times.
(Controller.prototype.handlePanelTransitionEnd):
(Controller.prototype.setPlaying):
(Controller.prototype.showControls):
(Controller.prototype.hideControls):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182263 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMedia controls should not always auto-hide
eric.carlson@apple.com [Wed, 1 Apr 2015 23:43:50 +0000 (23:43 +0000)]
Media controls should not always auto-hide
https://bugs.webkit.org/show_bug.cgi?id=143322

Reviewed by Dean Jackson.

* Modules/mediacontrols/mediaControlsApple.js:
(Controller): Intialize new properties.
(Controller.prototype.hideControls): Do nothing if controlsAlwaysVisible() returns true;
(Controller.prototype.controlsAlwaysVisible): New.
(Controller.prototype.controlsAreHidden): Consult controlsAlwaysVisible().
(Controller.prototype.currentPlaybackTargetIsWireless): Use new properties.
(Controller.prototype.updateWirelessTargetAvailable): Cache video.webkitCurrentPlaybackTargetIsWireless
    and video.webkitWirelessVideoPlaybackDisabled because we know when they change and
    use them frequently.
* Modules/mediacontrols/mediaControlsiOS.js:
(ControllerIOS.prototype.controlsAlwaysVisible): New.
* platform/graphics/MediaPlaybackTarget.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182262 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Win] Run test-webkitpy as part of EWS (just like we do on other platforms)
bfulgham@apple.com [Wed, 1 Apr 2015 23:43:17 +0000 (23:43 +0000)]
[Win] Run test-webkitpy as part of EWS (just like we do on other platforms)
https://bugs.webkit.org/show_bug.cgi?id=143321

Reviewed by Alexey Proskuryakov.

* EWSTools/start-queue-win.sh: Add commands to run 'test-webkitpy' during processing loop.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182261 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoInjected bundle messages should be at the page level.
enrica@apple.com [Wed, 1 Apr 2015 23:22:19 +0000 (23:22 +0000)]
Injected bundle messages should be at the page level.
https://bugs.webkit.org/show_bug.cgi?id=143283

Reviewed by Alexey Proskuryakov.

Source/WebKit2:

All the messages between the injected bundle and the UI
process should be page level messages instead of context.
The author of this patch is Tim Horton, I just finished it.

* UIProcess/API/C/WKPage.cpp:
(WKPageSetPageInjectedBundleClient):
* UIProcess/API/C/WKPage.h:
* UIProcess/API/C/WKPageInjectedBundleClient.h: Added.
* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _pageForTesting]):
* UIProcess/API/Cocoa/WKWebViewInternal.h:
* UIProcess/WebPageInjectedBundleClient.cpp: Added.
(WebKit::WebPageInjectedBundleClient::didReceiveMessageFromInjectedBundle):
(WebKit::WebPageInjectedBundleClient::didReceiveSynchronousMessageFromInjectedBundle):
* UIProcess/WebPageInjectedBundleClient.h: Added.
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::setInjectedBundleClient):
(WebKit::WebPageProxy::handleMessage):
(WebKit::WebPageProxy::handleSynchronousMessage):
* UIProcess/WebPageProxy.h:
(WebKit::WebPageProxy::injectedBundleClient):
* UIProcess/WebPageProxy.messages.in:
* UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::handleMessage):
* WebKit2.xcodeproj/project.pbxproj:
* CMakeLists.txt:
* WebProcess/InjectedBundle/API/c/WKBundlePage.cpp:
(WKBundlePagePostMessage):
(WKBundlePagePostSynchronousMessage):
* WebProcess/InjectedBundle/API/c/WKBundlePage.h:
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::postMessage):
(WebKit::WebPage::postSynchronousMessage):
* WebProcess/WebPage/WebPage.h:

Tools:

Updating WebKitTestRunner to use the new page level messaging.
The author of this patch is Tim Horton, I just finished it.

* WebKitTestRunner/InjectedBundle/EventSendingController.cpp:
(WTR::EventSendingController::mouseDown):
(WTR::EventSendingController::mouseUp):
(WTR::EventSendingController::mouseMoveTo):
(WTR::EventSendingController::leapForward):
(WTR::EventSendingController::scheduleAsynchronousClick):
(WTR::EventSendingController::keyDown):
(WTR::EventSendingController::scheduleAsynchronousKeyDown):
(WTR::EventSendingController::mouseScrollBy):
(WTR::EventSendingController::mouseScrollByWithWheelAndMomentumPhases):
(WTR::EventSendingController::continuousMouseScrollBy):
(WTR::EventSendingController::addTouchPoint):
(WTR::EventSendingController::updateTouchPoint):
(WTR::EventSendingController::setTouchModifier):
(WTR::EventSendingController::setTouchPointRadius):
(WTR::EventSendingController::touchStart):
(WTR::EventSendingController::touchMove):
(WTR::EventSendingController::touchEnd):
(WTR::EventSendingController::touchCancel):
(WTR::EventSendingController::clearTouchPoints):
(WTR::EventSendingController::releaseTouchPoint):
(WTR::EventSendingController::cancelTouchPoint):
* WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
(WTR::InjectedBundle::didReceiveMessage):
(WTR::InjectedBundle::didReceiveMessageToPage):
(WTR::InjectedBundle::done):
(WTR::InjectedBundle::outputText):
(WTR::InjectedBundle::postNewBeforeUnloadReturnValue):
(WTR::InjectedBundle::postAddChromeInputField):
(WTR::InjectedBundle::postRemoveChromeInputField):
(WTR::InjectedBundle::postFocusWebView):
(WTR::InjectedBundle::postSetBackingScaleFactor):
(WTR::InjectedBundle::postSetWindowIsKey):
(WTR::InjectedBundle::postSimulateWebNotificationClick):
(WTR::InjectedBundle::postSetAddsVisitedLinks):
(WTR::InjectedBundle::setGeolocationPermission):
(WTR::InjectedBundle::setMockGeolocationPosition):
(WTR::InjectedBundle::setMockGeolocationPositionUnavailableError):
(WTR::InjectedBundle::setUserMediaPermission):
(WTR::InjectedBundle::setCustomPolicyDelegate):
(WTR::InjectedBundle::setHidden):
(WTR::InjectedBundle::setCacheModel):
(WTR::InjectedBundle::shouldProcessWorkQueue):
(WTR::InjectedBundle::processWorkQueue):
(WTR::InjectedBundle::queueBackNavigation):
(WTR::InjectedBundle::queueForwardNavigation):
(WTR::InjectedBundle::queueLoad):
(WTR::InjectedBundle::queueLoadHTMLString):
(WTR::InjectedBundle::queueReload):
(WTR::InjectedBundle::queueLoadingScript):
(WTR::InjectedBundle::queueNonLoadingScript):
* WebKitTestRunner/InjectedBundle/TestRunner.cpp:
(WTR::TestRunner::setHandlesAuthenticationChallenges):
(WTR::TestRunner::setAuthenticationUsername):
(WTR::TestRunner::setAuthenticationPassword):
(WTR::TestRunner::secureEventInputIsEnabled):
(WTR::TestRunner::setBlockAllPlugins):
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::createWebViewWithOptions):
(WTR::TestController::resetStateToConsistentValues):
(WTR::TestController::didReceivePageMessageFromInjectedBundle):
(WTR::TestController::didReceiveSynchronousPageMessageFromInjectedBundle):
* WebKitTestRunner/TestController.h:
* WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::invoke):
(WTR::TestInvocation::didReceiveMessageFromInjectedBundle):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182260 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[iOS] WebContent should have access to the entirety of its caches directory
ap@apple.com [Wed, 1 Apr 2015 22:50:53 +0000 (22:50 +0000)]
[iOS] WebContent should have access to the entirety of its caches directory
https://bugs.webkit.org/show_bug.cgi?id=143312
rdar://problem/20314916

Reviewed by Oliver Hunt.

* NetworkProcess/cocoa/NetworkProcessCocoa.mm:
(WebKit::NetworkProcess::platformInitializeNetworkProcessCocoa):
* Shared/Network/NetworkProcessCreationParameters.cpp:
(WebKit::NetworkProcessCreationParameters::encode):
(WebKit::NetworkProcessCreationParameters::decode):
* Shared/Network/NetworkProcessCreationParameters.h:
* Shared/WebProcessCreationParameters.cpp:
(WebKit::WebProcessCreationParameters::encode):
(WebKit::WebProcessCreationParameters::decode):
* Shared/WebProcessCreationParameters.h:
* UIProcess/Cocoa/WebProcessPoolCocoa.mm:
(WebKit::WebProcessPool::networkingCachesDirectory):
(WebKit::WebProcessPool::webContentCachesDirectory):
(WebKit::WebProcessPool::openGLCacheDirectory): Deleted.
(WebKit::WebProcessPool::networkingHSTSDatabasePath): Deleted.
(WebKit::WebProcessPool::webContentHSTSDatabasePath): Deleted.
* UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::ensureNetworkProcess):
(WebKit::WebProcessPool::createNewWebProcess):
* UIProcess/WebProcessPool.h:
* WebProcess/cocoa/WebProcessCocoa.mm:
(WebKit::WebProcess::platformInitializeWebProcess):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182259 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agofast/events/mouse-cursor-image-set.html is timing out on Apple platforms
dino@apple.com [Wed, 1 Apr 2015 22:48:33 +0000 (22:48 +0000)]
fast/events/mouse-cursor-image-set.html is timing out on Apple platforms
https://bugs.webkit.org/show_bug.cgi?id=143323

The change r182247 has caused this test to timeout. Rather than rollout
the commit I'm skipping the test and will contact Yoav.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182258 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoREGRESSION (r182121): SVG animation macros cause warnings in MSVC
ap@apple.com [Wed, 1 Apr 2015 22:13:05 +0000 (22:13 +0000)]
REGRESSION (r182121): SVG animation macros cause warnings in MSVC
https://bugs.webkit.org/show_bug.cgi?id=143313

Reviewed by Darin Adler.

* svg/SVGExternalResourcesRequired.h: Undo the change that caused a type mismatch.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182256 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAnother attempt to fixi the iOS build after r182240.
eric.carlson@apple.com [Wed, 1 Apr 2015 21:26:30 +0000 (21:26 +0000)]
Another attempt to fixi the iOS build after r182240.

Reviewed by Dean Jackson.

* platform/graphics/MediaPlaybackTarget.h:
(WebCore::MediaPlaybackTarget::hasActiveRoute):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182255 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago015-04-01 Antti Koivisto <antti@apple.com>
antti@apple.com [Wed, 1 Apr 2015 21:01:24 +0000 (21:01 +0000)]
015-04-01  Antti Koivisto  <antti@apple.com>

Value assignment operator of Optional should be stricter
https://bugs.webkit.org/show_bug.cgi?id=143303

Reviewed by Andreas Kling.

struct Foo { mutable Optional<std::chrono::seconds> s; };
Foo foo, foo2;
foo = foo2;

doesn't build because it tries to use wrong operator=.

* wtf/Optional.h:

    Add enable_if test similar to std::experimental::optional.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182254 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoaccessibility/textarea-selected-text-range.html failing on Windows WK1 Release
dino@apple.com [Wed, 1 Apr 2015 19:55:50 +0000 (19:55 +0000)]
accessibility/textarea-selected-text-range.html failing on Windows WK1 Release
https://bugs.webkit.org/show_bug.cgi?id=143315

Removing clashing entry but changing the old entry to Skip.

* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182252 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix the iOS build after r182240.
eric.carlson@apple.com [Wed, 1 Apr 2015 19:49:47 +0000 (19:49 +0000)]
Fix the iOS build after r182240.

* platform/graphics/MediaPlaybackTarget.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182250 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoaccessibility/textarea-selected-text-range.html failing on Windows WK1 Release
dino@apple.com [Wed, 1 Apr 2015 19:33:56 +0000 (19:33 +0000)]
accessibility/textarea-selected-text-range.html failing on Windows WK1 Release
https://bugs.webkit.org/show_bug.cgi?id=143315

Skipping this on Windows.

* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182249 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r182174 and r182177.
commit-queue@webkit.org [Wed, 1 Apr 2015 19:20:14 +0000 (19:20 +0000)]
Unreviewed, rolling out r182174 and r182177.
https://bugs.webkit.org/show_bug.cgi?id=143314

Try to correct EWS failures (Requested by bfulgham on
#webkit).

Reverted changesets:

"Python SCM should be able to retrieve untracked files"
https://bugs.webkit.org/show_bug.cgi?id=143135
http://trac.webkit.org/changeset/182174

"[buildbot] clean-build script should remove untracked files
and revert local changes too"
https://bugs.webkit.org/show_bug.cgi?id=142400
http://trac.webkit.org/changeset/182177

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182248 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAsync loading of image resources
yoav@yoav.ws [Wed, 1 Apr 2015 19:02:35 +0000 (19:02 +0000)]
Async loading of image resources
https://bugs.webkit.org/show_bug.cgi?id=134488

Reviewed by Dean Jackson.

Source/WebCore:

This patch changes image loading to be asynchronous (per spec), in order
for it to be agnostic to property setting order when it comes to src, srcset and crossorigin,
as well as to enable future hooking of the <picture>-based selection logic on top of it.

Tests: fast/dom/HTMLImageElement/image-load-post-onload.html
fast/dom/HTMLImageElement/sizes/image-sizes-js-change-reverse.html

* WebCore.xcodeproj/project.pbxproj: Change MicroTask.h to be private.
* bindings/js/JSEventListener.cpp:
(WebCore::JSEventListener::handleEvent): Add a microtask checkpoint after event handling.
* bindings/js/ScheduledAction.cpp:
(WebCore::ScheduledAction::execute): Add a microtask checkpoint after running of a scheduled action.
* html/HTMLImageElement.cpp:
(WebCore::HTMLImageElement::didMoveToNewDocument): Add the old document to call of elementDidMoveToNewDocument.
* html/HTMLImageLoader.cpp:
(WebCore::HTMLImageLoader::notifyFinished): Avoid a crash when notifyFinished is called and there's no CachedImage.
* html/HTMLInputElement.cpp:
(WebCore::HTMLInputElement::didMoveToNewDocument): Add the old document to call of elementDidMoveToNewDocument.
* html/HTMLPlugInImageElement.cpp:
(WebCore::HTMLPlugInImageElement::didMoveToNewDocument): Add the old document to call of elementDidMoveToNewDocument.
* html/HTMLVideoElement.cpp:
(WebCore::HTMLVideoElement::didMoveToNewDocument): Add the old document to call of elementDidMoveToNewDocument.
* html/parser/HTMLDocumentParser.cpp:
(WebCore::HTMLDocumentParser::attemptToRunDeferredScriptsAndEnd): Add a microtask checkpoint.
* loader/ImageLoader.cpp: Move image resource loading to be asynchronous.
(WebCore::ImageLoader::ImageLoaderTask::create):
(WebCore::ImageLoader::ImageLoaderTask::run): Run the image loading microtask and called doUpdateFromElement.
(WebCore::ImageLoader::ImageLoaderTask::ImageLoaderTask):
(WebCore::ImageLoader::ImageLoader):
(WebCore::ImageLoader::doUpdateFromElement): This is split from the previous updateFromElement, and performs the actual resource loading.
(WebCore::ImageLoader::updateFromElement): Now only prepares the stage for the actual image resource loading.
(WebCore::ImageLoader::shouldLoadImmediately): If this returns true, the image resource is loaded immediately, without queueing a microtask.
(WebCore::ImageLoader::notifyFinished): Add asserts.
(WebCore::ImageLoader::elementDidMoveToNewDocument): Handle document load event counters decrementing and incrementing.
* loader/ImageLoader.h:
(WebCore::ImageLoader::imageComplete): Make sure that the image is complete only if there aren't any pending tasks.
(WebCore::ImageLoader::hasPendingActivity): Make sure that pending activity takes pending tasks into account.
(WebCore::ImageLoader::hasPendingTask): Getter to know if an ImageLoader has a pending task. (Used by the image loading microtask)
(WebCore::ImageLoader::createWeakPtr): Create a weakPtr to be used by the microtask, so it can called back the loader if it's still alive.
* loader/cache/CachedResourceLoader.cpp:
(WebCore::CachedResourceLoader::canRequest): Add a 'shouldBypassMainWorldContentSecurityPolicy' flag and act upon it.
(WebCore::CachedResourceLoader::requestImage): Add a 'shouldBypassMainWorldContentSecurityPolicy' flag.
(WebCore::CachedResourceLoader::requestResource): Add a 'shouldBypassMainWorldContentSecurityPolicy' flag.
* loader/cache/CachedResourceLoader.h:
(WebCore::CachedResourceLoader::canRequest): Add a 'shouldBypassMainWorldContentSecurityPolicy' flag and act upon it.
(WebCore::CachedResourceLoader::requestImage): Add a 'shouldBypassMainWorldContentSecurityPolicy' flag.
(WebCore::CachedResourceLoader::requestResource): Add a 'shouldBypassMainWorldContentSecurityPolicy' flag.
* svg/SVGImageElement.cpp:
(WebCore::SVGImageElement::didMoveToNewDocument): Add the old document to call of elementDidMoveToNewDocument.
* xml/parser/XMLDocumentParser.cpp:
(WebCore::XMLDocumentParser::end): Add a microtask checkpoint after XML finishes parsing.

LayoutTests:

* fast/dom/HTMLImageElement/image-load-post-onload-expected.txt: Added.
* fast/dom/HTMLImageElement/image-load-post-onload.html: Added.
* fast/dom/HTMLImageElement/image-loading-gc.html: Cosmetic changes.
* fast/dom/HTMLImageElement/sizes/image-sizes-js-change-reverse-expected.txt: Added.
* fast/dom/HTMLImageElement/sizes/image-sizes-js-change-reverse.html: Added.
* fast/dom/image-object.html: Cosmetic changes.
* http/tests/misc/image-blocked-src-change-expected.txt: Removed line numbers from console.
* http/tests/misc/image-blocked-src-no-change-expected.txt: Removed line numbers from console.
* http/tests/security/contentSecurityPolicy/report-blocked-file-uri-expected.txt: Removed line numbers from console.
* http/tests/security/frame-loading-via-document-write-expected.txt: Removed line numbers from console.
* http/tests/security/local-image-from-remote-expected.txt: Removed line numbers from console.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182247 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix Windows build.
simon.fraser@apple.com [Wed, 1 Apr 2015 19:00:22 +0000 (19:00 +0000)]
Fix Windows build.

* WebCoreSupport/WebChromeClient.h:
(WebChromeClient::numWheelEventHandlersChanged): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182246 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSmart magnification gesture sometimes shoots to the middle of the page
timothy_horton@apple.com [Wed, 1 Apr 2015 18:57:25 +0000 (18:57 +0000)]
Smart magnification gesture sometimes shoots to the middle of the page
https://bugs.webkit.org/show_bug.cgi?id=143296
<rdar://problem/18209280>

Reviewed by Dean Jackson.

* UIProcess/mac/ViewGestureControllerMac.mm:
(WebKit::ViewGestureController::didCollectGeometryForSmartMagnificationGesture):
Constrain the target rect to the viewport, and if it had overflowed the viewport,
scroll halfway towards the gesture origin.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182245 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed build fix when using content extensions performance reporting.
achristensen@apple.com [Wed, 1 Apr 2015 18:42:52 +0000 (18:42 +0000)]
Unreviewed build fix when using content extensions performance reporting.

* contentextensions/ContentExtensionCompiler.cpp:
(WebCore::ContentExtensions::compileRuleList):
Time how long it takes to build the DFAs.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182244 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoProgress towards CMake on Windows and Mac.
achristensen@apple.com [Wed, 1 Apr 2015 18:36:43 +0000 (18:36 +0000)]
Progress towards CMake on Windows and Mac.
https://bugs.webkit.org/show_bug.cgi?id=143293

Reviewed by Filip Pizlo.

.:

* CMakeLists.txt:
Set DERIVED_SOURCES_WTF_DIR for Windows.
* Source/CMakeLists.txt:
Don't compile bmalloc on Windows.
* Source/cmake/OptionsCommon.cmake:
Use the absolute path of the C preprocessor.
* Source/cmake/OptionsWinCairo.cmake:
Added needed definitions.
* Source/cmake/OptionsWindows.cmake:
Set some default values and removed support for old Visual Studio versions before /MP.
* Source/cmake/WebKitFS.cmake:
Make WTF DerivedSources directory.
* Source/cmake/WebKitMacros.cmake:
Added ADD_PRECOMPILED_HEADER macro based on
http://stackoverflow.com/questions/148570/using-pre-compiled-headers-with-cmake

Source/bmalloc:

* bmalloc/BAssert.h:
Removed ellipses from macros to appease Visual Studio.

Source/JavaScriptCore:

* CMakeLists.txt:
Enabled using assembly on Windows.
Replaced unix commands with CMake commands.
* PlatformMac.cmake:
Tell open source builders where to find unicode headers.

Source/ThirdParty/ANGLE:

* include/GLES2/gl2.h:
Temporarily disabled WebGL on Windows CMake builds.

Source/WebCore:

* PlatformMac.cmake:
Tell open source builders where to find unicode headers.
* PlatformWin.cmake:
Include PlatformWinCairo.cmake.
* PlatformWinCairo.cmake:
Added from bug 115944 by Patrick Gansterer.
* bindings/js/JSDOMStringListCustom.cpp:
* bindings/js/JSDOMWrapper.cpp:
* bindings/js/JSMessageChannelCustom.cpp:
* bindings/js/JSPopStateEventCustom.cpp:
* bindings/js/JSReadableStreamCustom.cpp:
* bindings/js/ReadableStreamJSSource.cpp:
* bindings/js/ScriptController.cpp:
* css/CSSComputedStyleDeclaration.cpp:
* dom/Attr.cpp:
* dom/CollectionIndexCache.cpp:
* platform/graphics/ANGLEWebKitBridge.h:
Fixed include quirks.

Source/WebKit:

* PlatformMac.cmake:
Fixed some include directories.

Source/WTF:

* wtf/CMakeLists.txt:
Don't use bmalloc on Windows yet.
* wtf/FeatureDefines.h:
Temporarily disabled WebGL on Windows CMake builds.
* wtf/PlatformMac.cmake:
Tell open source builders where to find unicode headers.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182243 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRemove scrolling tree dependency on wheel event handler counts, and use fast scrollin...
simon.fraser@apple.com [Wed, 1 Apr 2015 18:30:39 +0000 (18:30 +0000)]
Remove scrolling tree dependency on wheel event handler counts, and use fast scrolling even when there are wheel handlers
https://bugs.webkit.org/show_bug.cgi?id=143288
rdar://problem/16049624

Reviewed by Beth Dakin.

Remove the wheel event counting that Document does, and passes into the scrolling tree.
The ScrollingTree now just uses the non-fast scrollable region to determine when to
fast scroll on pages with wheel event handlers.

If a handler includes position:fixed renderers, we just cover the whole document
with the slow-scrolling region currently. This could be improved.

* dom/Document.cpp:
(WebCore::Document::Document):
(WebCore::Document::didBecomeCurrentDocumentInFrame):
(WebCore::Document::wheelEventHandlersChanged):
(WebCore::Document::didAddWheelEventHandler):
(WebCore::Document::didRemoveWheelEventHandler):
(WebCore::Document::wheelEventHandlerCount):
(WebCore::Document::touchEventHandlerCount):
(WebCore::Document::absoluteRegionForEventTargets): Changed to return a pair<Region, bool>
where the bool indicates whether any handler includes position:fixed content.
(WebCore::pageWheelEventHandlerCountChanged): Deleted.
(WebCore::wheelEventHandlerCountChanged): Deleted.
* dom/Document.h:
(WebCore::Document::wheelEventHandlerCount): Deleted.
* loader/EmptyClients.h:
* page/ChromeClient.h:
* page/DebugPageOverlays.cpp:
(WebCore::MouseWheelRegionOverlay::updateRegion):
* page/scrolling/AsyncScrollingCoordinator.cpp:
(WebCore::AsyncScrollingCoordinator::frameViewNonFastScrollableRegionChanged):
(WebCore::AsyncScrollingCoordinator::recomputeWheelEventHandlerCountForFrameView): Deleted.
* page/scrolling/AsyncScrollingCoordinator.h:
* page/scrolling/ScrollingCoordinator.cpp:
(WebCore::ScrollingCoordinator::computeNonFastScrollableRegion):
(WebCore::ScrollingCoordinator::frameViewRootLayerDidChange):
(WebCore::ScrollingCoordinator::computeCurrentWheelEventHandlerCount): Deleted.
(WebCore::ScrollingCoordinator::frameViewWheelEventHandlerCountChanged): Deleted.
* page/scrolling/ScrollingCoordinator.h:
(WebCore::ScrollingCoordinator::recomputeWheelEventHandlerCountForFrameView): Deleted.
* page/scrolling/ScrollingStateFrameScrollingNode.cpp:
(WebCore::ScrollingStateFrameScrollingNode::ScrollingStateFrameScrollingNode):
(WebCore::ScrollingStateFrameScrollingNode::setWheelEventHandlerCount): Deleted.
* page/scrolling/ScrollingStateFrameScrollingNode.h:
* page/scrolling/ScrollingTree.cpp:
(WebCore::ScrollingTree::ScrollingTree):
(WebCore::ScrollingTree::shouldHandleWheelEventSynchronously):
(WebCore::ScrollingTree::commitNewTreeState):
* page/scrolling/ScrollingTree.h:
* page/scrolling/mac/ScrollingCoordinatorMac.mm:
(WebCore::ScrollingCoordinatorMac::updateTiledScrollingIndicator):
* page/scrolling/mac/ScrollingTreeFrameScrollingNodeMac.mm:
(WebCore::ScrollingTreeFrameScrollingNodeMac::updateBeforeChildren):
(WebCore::logWheelEventHandlerCountChanged): Deleted.
* testing/Internals.cpp:
(WebCore::Internals::touchEventHandlerCount):

Source/WebKit/mac:
Remove scrolling tree dependency on wheel event handler counts, and use fast scrolling even when there are wheel handlers
https://bugs.webkit.org/show_bug.cgi?id=143288

Reviewed by Beth Dakin.

Remove the wheel event counting that Document does, and passes into the scrolling tree.
The ScrollingTree now just uses the non-fast scrollable region to determine when to
fast scroll on pages with wheel event handlers.

* WebCoreSupport/WebChromeClient.h:

Source/WebKit2:
Remove scrolling tree dependency on wheel event handler counts, and use fast scrolling even when there are wheel handlers
https://bugs.webkit.org/show_bug.cgi?id=143288
rdar://problem/16049624

Reviewed by Beth Dakin.

Remove the wheel event counting that Document does, and passes into the scrolling tree.
The ScrollingTree now just uses the non-fast scrollable region to determine when to
fast scroll on pages with wheel event handlers.

If a handler includes position:fixed renderers, we just cover the whole document
with the slow-scrolling region currently. This could be improved.

* Shared/Scrolling/RemoteScrollingCoordinatorTransaction.cpp:
(ArgumentCoder<ScrollingStateFrameScrollingNode>::encode):
(ArgumentCoder<ScrollingStateFrameScrollingNode>::decode):
(WebKit::RemoteScrollingTreeTextStream::dump):
* WebProcess/WebCoreSupport/WebChromeClient.cpp:
(WebKit::WebChromeClient::wheelEventHandlersChanged):
(WebKit::WebChromeClient::numWheelEventHandlersChanged): Deleted.
* WebProcess/WebCoreSupport/WebChromeClient.h:
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::WebPage):
(WebKit::WebPage::wheelEventHandlersChanged):
(WebKit::WebPage::recomputeShortCircuitHorizontalWheelEventsState):
(WebKit::WebPage::numWheelEventHandlersChanged): Deleted.
* WebProcess/WebPage/WebPage.h:

LayoutTests:
Make it possible to compute a region for elements on the page that have wheel event handlers
https://bugs.webkit.org/show_bug.cgi?id=142807

Reviewed by Beth Dakin.

Update results, since any handler with position:fixed now causes the region to cover the document.

* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-fixed-child-expected.txt:
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-inside-fixed-expected.txt:
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-on-fixed-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182242 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[New Block-Inside-Inline Model] Floats need to be allowed to intrude into anonymous...
hyatt@apple.com [Wed, 1 Apr 2015 18:10:32 +0000 (18:10 +0000)]
[New Block-Inside-Inline Model] Floats need to be allowed to intrude into anonymous inline-blocks.
https://bugs.webkit.org/show_bug.cgi?id=143307

Reviewed by Dean Jackson.

Source/WebCore:

Added fast/block/inside-inlines/basic-float-intrusion.html

* rendering/InlineFlowBox.cpp:
(WebCore::InlineFlowBox::placeBoxesInBlockDirection):
Make sure to exclude the margins of an anonymous inline-block when considering the height of the line.

* rendering/RenderBlock.cpp:
(WebCore::RenderBlock::lineHeight):
(WebCore::RenderBlock::baselinePosition):
Anonymous inline-block lines ignore line-height and just return the height of the block. For baseline position, the baseline
is considered to be at the bottom border edge of the anonymous inline-block.

* rendering/RenderBlockFlow.cpp:
(WebCore::RenderBlockFlow::rebuildFloatingObjectSetFromIntrudingFloats):
(WebCore::RenderBlockFlow::addIntrudingFloats):
* rendering/RenderBlockFlow.h:
Patch float intrusion to handle the case of anonymous inline-blocks. Instead of collecting the floats from our parent(), we
collect them from the containingBlock() in the anonymous inline-block case. Patch addIntrudingFloats to pass in the container
that acts as the "parent", since it is not the parent() in the anonymous inline-block case.

* rendering/RenderBlockLineLayout.cpp:
(WebCore::updateLogicalInlinePositions):
Make sure to use the full available width for content always in the anonymous inline-block case, since we don't avoid the
floats, but instead let them intrude into the inline-block.

(WebCore::RenderBlockFlow::computeInlineDirectionPositionsForLine):
Patch callers of updateLogicalInlinePositions to pass in the root line box that we're checking for.

(WebCore::RenderBlockFlow::layoutLineBoxes):
Defer the layout of anonymous inline-blocks, since they need to only lay out once the vertical position to lay them out
at is known.

* rendering/RenderBox.cpp:
(WebCore::RenderBox::createsNewFormattingContext):
(WebCore::RenderBox::avoidsFloats):
Anonymous inline-blocks do not establish a new block formatting context, nor do they avoid floats.

* rendering/line/BreakingContext.h:
(WebCore::BreakingContext::handleReplaced):
Add code to do a just-in-time layout of the anonymous inline-block at the time it is being examined (after the break before
it was handled).

LayoutTests:

* fast/block/inside-inlines/basic-float-intrusion-expected.html: Added.
* fast/block/inside-inlines/basic-float-intrusion.html: Added.
* fast/block/inside-inlines/new-model/basic-float-intrusion-expected.html: Added.
* fast/block/inside-inlines/new-model/basic-float-intrusion.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182241 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Mac] Do not include route button if element does not support target playback
eric.carlson@apple.com [Wed, 1 Apr 2015 18:07:16 +0000 (18:07 +0000)]
[Mac] Do not include route button if element does not support target playback
https://bugs.webkit.org/show_bug.cgi?id=143251

Reviewed by Jer Noble.

* Modules/mediacontrols/mediaControlsApple.js:
(Controller.prototype.currentPlaybackTargetIsWireless): Return false if target playback is disabled.
(Controller.prototype.updateWirelessTargetAvailable): Do not show the picker if target
    playback is disabled.
(Controller.prototype.handleWirelessPlaybackChange): Call updateWirelessTargetAvailable.

* dom/Document.cpp:
(WebCore::Document::showPlaybackTargetPicker): Update for Page::showPlaybackTargetPicker change.
(WebCore::Document::didChoosePlaybackTarget): Notify clients in vector-order, lower level code
    now tracks which one requested the picker.

* html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::loadResource): Don't call applyMediaPlayerRestrictions, it is done
    in mediaPlayerEngineFailedToLoad.
(WebCore::HTMLMediaElement::setReadyState): Enqueue a target availability event if there are
    listeners readyState reaches HAVE_METADATA so controls are setup correctly.
(WebCore::HTMLMediaElement::mediaPlayerEngineFailedToLoad): Call applyMediaPlayerRestrictions.
(WebCore::HTMLMediaElement::clearMediaPlayer): Enqueue a target availability event in so controls
    have a chance to hide the picker.
(WebCore::HTMLMediaElement::webkitCurrentPlaybackTargetIsSupported): New, passthrough to the
    media engine.
(WebCore::HTMLMediaElement::canPlayToWirelessPlaybackTarget): Ditto.
(WebCore::HTMLMediaElement::startPlayingToPlaybackTarget): Ditto.
(WebCore::HTMLMediaElement::stopPlayingToPlaybackTarget): Ditto.
* html/HTMLMediaElement.h:
* html/HTMLMediaElement.idl:

* html/HTMLMediaSession.cpp:
(WebCore::HTMLMediaSession::HTMLMediaSession): Initialize m_playbackTarget.
(WebCore::HTMLMediaSession::currentPlaybackTargetIsSupported): New.
(WebCore::HTMLMediaSession::showPlaybackTargetPicker): Pull logic from showingPlaybackTargetPickerPermitted
    inline. Don't refuse to show a picker if the element doesn't support target playback, it is up
    to script to decide which elements can display a picker.
(WebCore::HTMLMediaSession::wirelessVideoPlaybackDisabled): Return true if there is no player.
(WebCore::HTMLMediaSession::didChoosePlaybackTarget): Call startPlayingToPlaybackTarget or
    stopPlayingToPlaybackTarget because setWirelessPlaybackTarget doesn't apply the target.
(WebCore::HTMLMediaSession::externalOutputDeviceAvailableDidChange): Add logging.
(WebCore::HTMLMediaSession::showingPlaybackTargetPickerPermitted): Deleted.
* html/HTMLMediaSession.h:

* page/Page.cpp:
(WebCore::Page::Page): Initialize m_playbackTarget.
(WebCore::Page::showPlaybackTargetPicker): Don't set m_documentRequestingPlaybackTargetPicker.
(WebCore::Page::didChoosePlaybackTarget): Notify Pages in vector-order, lower level code
    now tracks which one requested the picker.
(WebCore::Page::configurePlaybackTargetMonitoring): Don't track m_requiresPlaybackTargetMonitoring, it
    is too easy for it to get out of sync with the UI process state.
* page/Page.h:

* platform/audio/MediaSession.cpp:
(WebCore::MediaSession::canPlayToWirelessPlaybackTarget): New, client passthrough.
(WebCore::MediaSession::startPlayingToPlaybackTarget): Ditto.
(WebCore::MediaSession::stopPlayingToPlaybackTarget): Ditto.
* platform/audio/MediaSession.h:
(WebCore::MediaSessionClient::canPlayToWirelessPlaybackTarget):
(WebCore::MediaSessionClient::startPlayingToPlaybackTarget):
(WebCore::MediaSessionClient::stopPlayingToPlaybackTarget):

* platform/audio/MediaSessionManager.cpp:
(WebCore::MediaSessionManager::sessionWillBeginPlayback): Call startPlayingToPlaybackTarget &
    stopPlayingToPlaybackTarget as necessary.
* platform/audio/MediaSessionManager.h:

* platform/graphics/MediaPlaybackTarget.h:
(WebCore::MediaPlaybackTarget::hasActiveRoute): New.
* platform/graphics/MediaPlaybackTargetPickerClient.h:
* platform/graphics/MediaPlayer.cpp:
(WebCore::MediaPlayer::isCurrentPlaybackTargetSupported): New, engine passthrough.
(WebCore::MediaPlayer::canPlayToWirelessPlaybackTarget): Ditto.
(WebCore::MediaPlayer::startPlayingToPlaybackTarget): Ditto.
(WebCore::MediaPlayer::stopPlayingToPlaybackTarget): Ditto.
* platform/graphics/MediaPlayer.h:
* platform/graphics/MediaPlayerPrivate.h:
(WebCore::MediaPlayerPrivateInterface::isCurrentPlaybackTargetSupported):
(WebCore::MediaPlayerPrivateInterface::wirelessVideoPlaybackDisabled):
(WebCore::MediaPlayerPrivateInterface::canPlayToWirelessPlaybackTarget):
(WebCore::MediaPlayerPrivateInterface::startPlayingToPlaybackTarget):
(WebCore::MediaPlayerPrivateInterface::stopPlayingToPlaybackTarget):

* platform/graphics/avfoundation/MediaPlaybackTargetMac.mm:
(WebCore::MediaPlaybackTarget::hasActiveRoute): New.

* platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.h:
(WebCore::MediaPlayerPrivateAVFoundationObjC::canPlayToWirelessPlaybackTarget):
* platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
(WebCore::MediaPlayerPrivateAVFoundationObjC::cancelLoad): Use playerKVOProperties. Drive-by
    code cleanup.
(WebCore::MediaPlayerPrivateAVFoundationObjC::createAVPlayer): Use playerKVOProperties.
(WebCore::MediaPlayerPrivateAVFoundationObjC::wirelessPlaybackTargetName): Implement for Mac.
(WebCore::MediaPlayerPrivateAVFoundationObjC::setWirelessVideoPlaybackDisabled): Delay callbacks
    while setting AVPlayer property.
(WebCore::MediaPlayerPrivateAVFoundationObjC::setWirelessPlaybackTarget): Add logging. Don't set
    the AVPlayer outputContext immediately.
(WebCore::MediaPlayerPrivateAVFoundationObjC::startPlayingToPlaybackTarget): New.
(WebCore::MediaPlayerPrivateAVFoundationObjC::stopPlayingToPlaybackTarget): New.
(WebCore::playerKVOProperties): New.
(-[WebCoreAVFMovieObserver observeValueForKeyPath:ofObject:change:context:]): Notify when
    allowsExternalPlayback changes.

* platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaSourceAVFObjC.h:
* platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaSourceAVFObjC.mm:
(WebCore::MediaPlayerPrivateMediaSourceAVFObjC::isCurrentPlaybackTargetSupported): New.
(WebCore::MediaPlayerPrivateMediaSourceAVFObjC::setWirelessPlaybackTarget): Ditto.
(WebCore::MediaPlayerPrivateMediaSourceAVFObjC::togglePlayingToPlaybackTarget): Ditto.
(WebCore::MediaPlayerPrivateMediaSourceAVFObjC::startPlayingToPlaybackTarget): Ditto.
(WebCore::MediaPlayerPrivateMediaSourceAVFObjC::stopPlayingToPlaybackTarget): Ditto.

* platform/graphics/mac/MediaPlayerPrivateQTKit.h:
* platform/graphics/mac/MediaPlayerPrivateQTKit.mm:
(WebCore::MediaPlayerPrivateQTKit::isCurrentPlaybackTargetSupported): Ditto.
(WebCore::MediaPlayerPrivateQTKit::setWirelessPlaybackTarget): Ditto.
(WebCore::MediaPlayerPrivateQTKit::togglePlayingToPlaybackTarget): Ditto.
(WebCore::MediaPlayerPrivateQTKit::startPlayingToPlaybackTarget): Ditto.
(WebCore::MediaPlayerPrivateQTKit::stopPlayingToPlaybackTarget): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182240 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Content Extensions] Properly handle regexes that match everything.
achristensen@apple.com [Wed, 1 Apr 2015 17:48:57 +0000 (17:48 +0000)]
[Content Extensions] Properly handle regexes that match everything.
https://bugs.webkit.org/show_bug.cgi?id=143281

Reviewed by Benjamin Poulain.

Source/WebCore:

* contentextensions/ContentExtensionCompiler.cpp:
(WebCore::ContentExtensions::compileRuleList):
Make sure we always have at least one NFA for rule lists with triggers that all match everything.
* contentextensions/DFABytecodeInterpreter.cpp:
(WebCore::ContentExtensions::DFABytecodeInterpreter::interpretAppendAction):
(WebCore::ContentExtensions::DFABytecodeInterpreter::interpretTestFlagsAndAppendAction):
Add helper functions to reduce duplicate code.
(WebCore::ContentExtensions::DFABytecodeInterpreter::actionsFromDFARoot):
Get all actions without flags from the DFA root.
(WebCore::ContentExtensions::DFABytecodeInterpreter::interpret):
Remove branches when interpreting.
* contentextensions/DFABytecodeInterpreter.h:
Add helper functions to reduce duplicate code.

Tools:

* TestWebKitAPI/Tests/WebCore/ContentExtensions.cpp:
(TestWebKitAPI::testRequest):
(TestWebKitAPI::makeBackend):
(TestWebKitAPI::TEST_F):
Removed duplicate code and added test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182239 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAddressing post-review comments.
mmaxfield@apple.com [Wed, 1 Apr 2015 17:43:21 +0000 (17:43 +0000)]
Addressing post-review comments.
https://bugs.webkit.org/show_bug.cgi?id=143292

Unreviewed.

* platform/graphics/FontCascade.cpp:
(WebCore::FontCascade::leadingExpansionOpportunity):
(WebCore::FontCascade::trailingExpansionOpportunity):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182238 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WinCairo] Link errors, unresolved texture mapper methods.
achristensen@apple.com [Wed, 1 Apr 2015 17:32:43 +0000 (17:32 +0000)]
[WinCairo] Link errors, unresolved texture mapper methods.
https://bugs.webkit.org/show_bug.cgi?id=143308

Patch by Per Arne Vollan <peavo@outlook.com> on 2015-04-01
Reviewed by Alex Christensen.

New texture mapper files should be included in the project.

* WebCore.vcxproj/WebCore.vcxproj:
* WebCore.vcxproj/WebCore.vcxproj.filters:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182237 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSupport forcing expansion opportunities at the beginning and ending of a run
mmaxfield@apple.com [Wed, 1 Apr 2015 16:58:20 +0000 (16:58 +0000)]
Support forcing expansion opportunities at the beginning and ending of a run
https://bugs.webkit.org/show_bug.cgi?id=143292

Reviewed by David Hyatt.

Migrate ExpansionBehavior to an enum comprised of two values: one for leading behavior
and one for trailing behavior. Each of these two values can take on a value of
"force," "forbid," and neither. All calls that interact with justification are
migrated to use this new structure.

Note that in this terminology, "leading" and "trailing" are with respect to the order
that elements get laid out in a line. Therefore, leading is always on the left, since
lines get laid out that way regardless of their bidi attributes.

This is getting ready for https://bugs.webkit.org/show_bug.cgi?id=142608

No new tests because there is no behavior change.

* platform/graphics/FontCascade.cpp:
(WebCore::FontCascade::expansionOpportunityCountInternal): Migrate to these new
values.
(WebCore::FontCascade::expansionOpportunityCount): Ditto.
(WebCore::FontCascade::leadingExpansionOpportunity): Returns whether one is present
or not.
(WebCore::FontCascade::trailingExpansionOpportunity): Ditto.
* platform/graphics/FontCascade.h:
* platform/graphics/GlyphBuffer.h: New leading expansion field.
(WebCore::GlyphBuffer::setLeadingExpansion):
(WebCore::GlyphBuffer::leadingExpansion):
* platform/graphics/TextRun.h: m_expansionBehavior needs more bits.
(WebCore::TextRun::expansionBehavior):
(WebCore::TextRun::allowsLeadingExpansion): Deleted.
(WebCore::TextRun::allowsTrailingExpansion): Deleted.
* platform/graphics/WidthIterator.cpp: Update to support new type.
(WebCore::WidthIterator::WidthIterator):
(WebCore::expansionLocation): Where should we insert expansions?
(WebCore::WidthIterator::advanceInternal): Use expansionLocation()
* platform/graphics/cocoa/FontCascadeCocoa.mm:
(WebCore::FontCascade::adjustSelectionRectForComplexText):
(WebCore::FontCascade::getGlyphsAndAdvancesForComplexText):
* platform/graphics/mac/ComplexTextController.cpp: Same as WidthIterator
(WebCore::ComplexTextController::ComplexTextController):
(WebCore::ComplexTextController::advance):
(WebCore::expansionLocation):
(WebCore::ComplexTextController::adjustGlyphsAndAdvances):
* platform/graphics/mac/ComplexTextController.h:
(WebCore::ComplexTextController::leadingExpansion):
* platform/text/TextFlags.h: Add new enum values
* rendering/InlineBox.h: Update to include new values.
(WebCore::InlineBox::InlineBoxBitfields::InlineBoxBitfields):
(WebCore::InlineBox::canHaveTrailingExpansion):
(WebCore::InlineBox::setCanHaveTrailingExpansion):
(WebCore::InlineBox::setForceTrailingExpansion):
(WebCore::InlineBox::forceTrailingExpansion):
(WebCore::InlineBox::setForceLeadingExpansion):
(WebCore::InlineBox::forceLeadingExpansion):
* rendering/InlineTextBox.h:
* rendering/RenderBlockLineLayout.cpp: Update to use new FontCascade signatures.
(WebCore::expansionBehaviorForInlineTextBox):
(WebCore::applyExpansionBehavior):
(WebCore::RenderBlockFlow::computeInlineDirectionPositionsForSegment):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182236 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoLots of time spent querying table cell borders, when there are none.
zalan@apple.com [Wed, 1 Apr 2015 15:56:34 +0000 (15:56 +0000)]
Lots of time spent querying table cell borders, when there are none.
https://bugs.webkit.org/show_bug.cgi?id=143277

Reviewed by Simon Fraser.

This patch speeds up collapsed border queries by using existing collapsed border
cache to calculate repaint rects and by introducing a fast path for zero width collapsed borders.

It reduces the number of calls to recompute collapsed borders from 36 000 to 1 600, while loading a page with a table of 400 rows (1 cell per row).
When scrolling the same page all the way down to the bottom, the number of calls to recompute collapsed borders falls from 290 000 to 0.

Covered by existing tests.

* rendering/RenderTable.cpp:
(WebCore::RenderTable::styleDidChange): This moves invalidation time from RenderTable::layout() to styleDidChange().
(WebCore::RenderTable::invalidateCollapsedBorders):
(WebCore::RenderTable::recalcCollapsedBorders):
* rendering/RenderTable.h:
(WebCore::RenderTable::collapsedBordersAreValid):
(WebCore::RenderTable::invalidateCollapsedBorders): Deleted.
* rendering/RenderTableCell.cpp:
(WebCore::RenderTableCell::RenderTableCell):
(WebCore::RenderTableCell::willBeRemovedFromTree): Invalidate caches so that when repaint rect is calculated, we don't end up using stale values.
(WebCore::RenderTableCell::styleDidChange): Same as willBeRemovedFromTree.
(WebCore::RenderTableCell::collapsedStartBorder): Check if collapsed border is zero -also query cache.
(WebCore::RenderTableCell::collapsedEndBorder):
(WebCore::RenderTableCell::collapsedBeforeBorder):
(WebCore::RenderTableCell::collapsedAfterBorder):
(WebCore::RenderTableCell::cachedCollapsedLeftBorder):
(WebCore::RenderTableCell::cachedCollapsedRightBorder):
(WebCore::RenderTableCell::cachedCollapsedTopBorder):
(WebCore::RenderTableCell::cachedCollapsedBottomBorder):
(WebCore::RenderTableCell::paintCollapsedBorders):
(WebCore::RenderTableCell::cellAtLeft): Deleted.
(WebCore::RenderTableCell::cellAtRight): Deleted.
(WebCore::RenderTableCell::cellAtTop): Deleted.
(WebCore::RenderTableCell::cellAtBottom): Deleted.
* rendering/RenderTableCell.h:
(WebCore::RenderTableCell::invalidateHasEmptyCollapsedBorders):
* rendering/RenderTableCol.cpp:
(WebCore::RenderTableCol::styleDidChange):
* rendering/RenderTableRow.cpp:
(WebCore::RenderTableRow::styleDidChange):
(WebCore::RenderTableRow::addChild):
* rendering/RenderTableSection.cpp:
(WebCore::RenderTableSection::styleDidChange):
(WebCore::RenderTableSection::clearCachedCollapsedBorders): This is just an extra safety to invalidate collapsed border cache. This is always
called together with RenderTable::invalidateCollapsedBorders() -and that should prevent the RenderCells to use the cache.
(WebCore::RenderTableSection::removeCachedCollapsedBorders):
(WebCore::RenderTableSection::setCachedCollapsedBorder):
(WebCore::RenderTableSection::cachedCollapsedBorder):
* rendering/RenderTableSection.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182235 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoVersioning.
bshafiei@apple.com [Wed, 1 Apr 2015 15:39:37 +0000 (15:39 +0000)]
Versioning.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182233 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[CMake][GTK] Use the right variable to include the Wayland headers.
clopez@igalia.com [Wed, 1 Apr 2015 14:30:18 +0000 (14:30 +0000)]
[CMake][GTK] Use the right variable to include the Wayland headers.
https://bugs.webkit.org/show_bug.cgi?id=143304

Reviewed by Carlos Garcia Campos.

No new tests, no behavior changes.

* PlatformGTK.cmake: Use the right variable WAYLAND_INCLUDE_DIRS.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182232 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed EFL gardening. Remove warning message in layoutTest.
commit-queue@webkit.org [Wed, 1 Apr 2015 13:39:43 +0000 (13:39 +0000)]
Unreviewed EFL gardening. Remove warning message in layoutTest.
https://bugs.webkit.org/show_bug.cgi?id=143268

Patch by Hunseop Jeong <hs85.jeong@samsung.com> on 2015-04-01

* platform/efl/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182231 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUse colored diagnostics when building with cmake + ninja + clang
yoon@igalia.com [Wed, 1 Apr 2015 11:15:01 +0000 (11:15 +0000)]
Use colored diagnostics when building with cmake + ninja + clang
https://bugs.webkit.org/show_bug.cgi?id=143297

Reviewed by Žan Doberšek.

Because that ninja sets subprocess stdout/stderr to a pipe, clang
disables colored output.
This patch forces clang to use colored diagnostics when we are using
the ninja.

* Source/cmake/OptionsCommon.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182230 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GStreamer] Move synchronous message handler to MediaPlayerPrivateGStreamerBase
philn@webkit.org [Wed, 1 Apr 2015 10:46:19 +0000 (10:46 +0000)]
[GStreamer] Move synchronous message handler to MediaPlayerPrivateGStreamerBase
https://bugs.webkit.org/show_bug.cgi?id=143218

Reviewed by Carlos Garcia Campos.

Moved the pipeline pointer to the base class, sub-classes should
now invoke setPipeline() when loading a resource. The base class
now takes care of configuring the GL shared context, this is more
logical because video rendering is also managed by this class.

No new tests, this is only a code refactoring.

* platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:
(WebCore::MediaPlayerPrivateGStreamer::~MediaPlayerPrivateGStreamer):
(WebCore::MediaPlayerPrivateGStreamer::load):
(WebCore::MediaPlayerPrivateGStreamer::playbackPosition):
(WebCore::MediaPlayerPrivateGStreamer::changePipelineState):
(WebCore::MediaPlayerPrivateGStreamer::pause):
(WebCore::MediaPlayerPrivateGStreamer::duration):
(WebCore::MediaPlayerPrivateGStreamer::currentTime):
(WebCore::MediaPlayerPrivateGStreamer::seek):
(WebCore::MediaPlayerPrivateGStreamer::doSeek):
(WebCore::MediaPlayerPrivateGStreamer::updatePlaybackRate):
(WebCore::MediaPlayerPrivateGStreamer::paused):
(WebCore::MediaPlayerPrivateGStreamer::notifyPlayerOfVideo):
(WebCore::MediaPlayerPrivateGStreamer::notifyPlayerOfAudio):
(WebCore::MediaPlayerPrivateGStreamer::notifyPlayerOfText):
(WebCore::MediaPlayerPrivateGStreamer::setRate):
(WebCore::MediaPlayerPrivateGStreamer::buffered):
(WebCore::MediaPlayerPrivateGStreamer::handleMessage):
(WebCore::MediaPlayerPrivateGStreamer::fillTimerFired):
(WebCore::MediaPlayerPrivateGStreamer::didLoadingProgress):
(WebCore::MediaPlayerPrivateGStreamer::sourceChanged):
(WebCore::MediaPlayerPrivateGStreamer::cancelLoad):
(WebCore::MediaPlayerPrivateGStreamer::asyncStateChangeDone):
(WebCore::MediaPlayerPrivateGStreamer::updateStates):
(WebCore::MediaPlayerPrivateGStreamer::loadNextLocation):
(WebCore::MediaPlayerPrivateGStreamer::cacheDuration):
(WebCore::MediaPlayerPrivateGStreamer::setDownloadBuffering):
(WebCore::MediaPlayerPrivateGStreamer::audioSink):
(WebCore::MediaPlayerPrivateGStreamer::createGSTPlayBin):
(WebCore::MediaPlayerPrivateGStreamer::simulateAudioInterruption):
(WebCore::mediaPlayerPrivateSyncMessageCallback): Deleted.
(WebCore::MediaPlayerPrivateGStreamer::handleSyncMessage): Deleted.
(WebCore::MediaPlayerPrivateGStreamer::ensureGstGLContext): Deleted.
* platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.h:
* platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.cpp:
(WebCore::mediaPlayerPrivateNeedContextMessageCallback):
(WebCore::MediaPlayerPrivateGStreamerBase::~MediaPlayerPrivateGStreamerBase):
(WebCore::MediaPlayerPrivateGStreamerBase::setPipeline):
(WebCore::MediaPlayerPrivateGStreamerBase::handleNeedContextMessage):
(WebCore::MediaPlayerPrivateGStreamerBase::ensureGstGLContext):
* platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182229 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[CMake] Make gtest/CMakeLists.txt more readable
commit-queue@webkit.org [Wed, 1 Apr 2015 10:38:06 +0000 (10:38 +0000)]
[CMake] Make gtest/CMakeLists.txt more readable
https://bugs.webkit.org/show_bug.cgi?id=143295

Patch by Gyuyoung Kim <gyuyoung.kim@webkit.org> on 2015-04-01
Reviewed by Csaba Osztrogonác.

gtest/CMakeLists.txt has used add_library() and include_directory() directly.
This patch adds GTEST_INCLUDE_DIRECTORIES and GTEST_SOURCES, and use it.

* gtest/CMakeLists.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182228 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GStreamer] GstGL support in the video sink
philn@webkit.org [Wed, 1 Apr 2015 10:19:03 +0000 (10:19 +0000)]
[GStreamer] GstGL support in the video sink
https://bugs.webkit.org/show_bug.cgi?id=138562

Use GStreamer's glimagesink for video rendering instead of our
custom video sink if a recent-enough version of GstGL is found
during the build. When glimagesink is used it passes a texture to
the media player which then wraps it inside a TextureMapper
texture later used for actual rendering.

Using this new code path will allow us to remove our custom sink
entirely in the long term.

Patch by Philippe Normand <philn@igalia.com> on 2015-03-24
Reviewed by Carlos Garcia Campos.

No new test, existing media tests cover video rendering already.

* platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.cpp:
(WebCore::mediaPlayerPrivateDrawCallback): client-draw signal callback.
(WebCore::MediaPlayerPrivateGStreamerBase::MediaPlayerPrivateGStreamerBase):
(WebCore::MediaPlayerPrivateGStreamerBase::~MediaPlayerPrivateGStreamerBase):
(WebCore::MediaPlayerPrivateGStreamerBase::naturalSize): Better
check for sample validity.
(WebCore::MediaPlayerPrivateGStreamerBase::updateTexture): Ditto.
(WebCore::MediaPlayerPrivateGStreamerBase::triggerRepaint): Trigger a
repaint and wait its completion.
(WebCore::MediaPlayerPrivateGStreamerBase::paintToTextureMapper):
Extract texture ID and video dimensions from the sample and do actual paint.
(WebCore::MediaPlayerPrivateGStreamerBase::createVideoSink): Use
glimagesink if GstGL is available.
* platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182227 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoIteratorClose should be called when jumping over the target for-of loop
utatane.tea@gmail.com [Wed, 1 Apr 2015 09:36:15 +0000 (09:36 +0000)]
IteratorClose should be called when jumping over the target for-of loop
https://bugs.webkit.org/show_bug.cgi?id=143140

Reviewed by Geoffrey Garen.

This patch fixes labeled break/continue behaviors with for-of and iterators.

1. Support IteratorClose beyond multiple loop contexts
Previously, IteratorClose is only executed in for-of's breakTarget().
However, this misses IteratorClose execution when statement roll-ups multiple control flow contexts.
For example,
outer: for (var e1 of outer) {
    inner: for (var e2 of inner) {
        break outer;
    }
}
In this case, return method of inner should be called.
We leverage the existing system for `finally` to execute inner.return method correctly.
Leveraging `finally` system fixes `break`, `continue` and `return` cases.
`throw` case is already supported by emitting try-catch handlers in for-of.

2. Incorrect LabelScope creation is done in ForOfNode
ForOfNode creates duplicated LabelScope.
It causes infinite loop when executing the following program that contains
explicitly labeled for-of loop.
For example,
inner: for (var elm of array) {
    continue inner;
}

* bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::pushFinallyContext):
(JSC::BytecodeGenerator::pushIteratorCloseContext):
(JSC::BytecodeGenerator::popFinallyContext):
(JSC::BytecodeGenerator::popIteratorCloseContext):
(JSC::BytecodeGenerator::emitComplexPopScopes):
(JSC::BytecodeGenerator::emitEnumeration):
(JSC::BytecodeGenerator::emitIteratorClose):
* bytecompiler/BytecodeGenerator.h:
* bytecompiler/NodesCodegen.cpp:
(JSC::ForOfNode::emitBytecode):
* tests/stress/iterator-return-beyond-multiple-iteration-scopes.js: Added.
(createIterator.iterator.return):
(createIterator):
* tests/stress/raise-error-in-iterator-close.js: Added.
(createIterator.iterator.return):
(createIterator):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182226 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[ES6] Implement Symbol.unscopables
utatane.tea@gmail.com [Wed, 1 Apr 2015 09:03:10 +0000 (09:03 +0000)]
[ES6] Implement Symbol.unscopables
https://bugs.webkit.org/show_bug.cgi?id=142829

Reviewed by Geoffrey Garen.

This patch introduces Symbol.unscopables functionality.
In ES6, some generic names (like keys, values) are introduced
as Array's method name. And this breaks the web since some web sites
use like the following code.

var values = ...;
with (array) {
    values;  // This values is trapped by array's method "values".
}

To fix this, Symbol.unscopables introduces blacklist
for with scope's trapping. When resolving scope,
if name is found in the target scope and the target scope is with scope,
we check Symbol.unscopables object to filter generic names.

This functionality is only active for with scopes.
Global scope does not have unscopables functionality.

And since
1) op_resolve_scope for with scope always return Dynamic resolve type,
2) in that case, JSScope::resolve is always used in JIT and LLInt,
3) the code which contains op_resolve_scope that returns Dynamic cannot be compiled with DFG and FTL,
to implement this functionality, we just change JSScope::resolve and no need to change JIT code.
So performance regression is only visible in Dynamic resolving case, and it is already much slow.

* runtime/ArrayPrototype.cpp:
(JSC::ArrayPrototype::finishCreation):
* runtime/CommonIdentifiers.h:
* runtime/JSGlobalObject.h:
(JSC::JSGlobalObject::runtimeFlags):
* runtime/JSScope.cpp:
(JSC::isUnscopable):
(JSC::JSScope::resolve):
* runtime/JSScope.h:
(JSC::ScopeChainIterator::scope):
* tests/stress/global-environment-does-not-trap-unscopables.js: Added.
(test):
* tests/stress/unscopables.js: Added.
(test):
(.):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182225 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agonullptr cleanup in StyleResolver
commit-queue@webkit.org [Wed, 1 Apr 2015 07:24:00 +0000 (07:24 +0000)]
nullptr cleanup in StyleResolver
https://bugs.webkit.org/show_bug.cgi?id=143260

Patch by Joonghun Park <jh718.park@samsung.com> on 2015-04-01
Reviewed by Ryosuke Niwa.

No new tests, no behavior changes

* css/StyleResolver.cpp:
(WebCore::StyleResolver::State::initForStyleResolve):
(WebCore::StyleResolver::locateCousinList):
(WebCore::StyleResolver::locateSharedStyle):
(WebCore::StyleResolver::keyframeStylesForAnimation):
(WebCore::StyleResolver::pseudoStyleForElement):
(WebCore::StyleResolver::pseudoStyleRulesForElement):
(WebCore::StyleResolver::findFromMatchedPropertiesCache):
(WebCore::StyleResolver::applyMatchedProperties):
(WebCore::StyleResolver::applyPropertyToStyle):
(WebCore::StyleResolver::MatchedProperties::MatchedProperties):
* css/StyleResolver.h:
(WebCore::PseudoStyleRequest::PseudoStyleRequest):
(WebCore::StyleResolverParentPusher::StyleResolverParentPusher):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182224 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoREGRESSION (r179820): accessibility/textarea-selected-text-range.html is flaky
ap@apple.com [Wed, 1 Apr 2015 07:07:12 +0000 (07:07 +0000)]
REGRESSION (r179820): accessibility/textarea-selected-text-range.html is flaky
https://bugs.webkit.org/show_bug.cgi?id=142976

Reviewed by Chris Fleizach.

* platform/mac/TestExpectations: Removed an expectation.

* accessibility/textarea-selected-text-range-expected.txt:
* accessibility/textarea-selected-text-range.html:
Don't race a timer with an internal dispatch_async.

* platform/mac/accessibility/select-element-selection-with-optgroups-expected.txt:
* platform/mac/accessibility/select-element-selection-with-optgroups.html:
This test had the same issue.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182223 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix the build after <http://trac.webkit.org/changeset/182205>.
ap@apple.com [Wed, 1 Apr 2015 06:18:30 +0000 (06:18 +0000)]
Fix the build after <trac.webkit.org/changeset/182205>.

* wtf/text/StringImpl.h: (WTF::StringImpl::extractFoldedStringInSymbol):
A function cannot be both exported and inline.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182222 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: console.error(object) has double disclosure triangles
nvasilyev@apple.com [Wed, 1 Apr 2015 06:16:48 +0000 (06:16 +0000)]
Web Inspector: console.error(object) has double disclosure triangles
https://bugs.webkit.org/show_bug.cgi?id=142069

Use inline-block instead of "float: left" or "position: absolute" to
make layout more predictable and easier to understand.

Reviewed by Timothy Hatcher.

* UserInterface/Views/LogContentView.css:
(.outline-disclosure, .outline-disclosure ol):
(.outline-disclosure > li.parent):
(.outline-disclosure > li.parent::before):
(.outline-disclosure li.parent): Deleted.
(.outline-disclosure li.parent::before): Deleted.
* UserInterface/Views/ObjectTreePropertyTreeElement.css:
(.object-tree-property > .titles):
(.object-tree-property > .disclosure-button):
(.object-tree-property > .icon):
* UserInterface/Views/ObjectTreeView.css:
(.object-tree > :matches(.title, .object-preview)::before):
(.object-tree:not(.lossless-preview) > :matches(.title, .object-preview)):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182221 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: REGRESSION: Filtering no longer works
jonowells@apple.com [Wed, 1 Apr 2015 05:00:02 +0000 (05:00 +0000)]
Web Inspector: REGRESSION: Filtering no longer works
https://bugs.webkit.org/show_bug.cgi?id=143099

Reviewed by Brian Burg.

Add a check for an empty array of filter functions and return true in this case to ensure text
and timeline filters work correctly.

* UserInterface/Views/NavigationSidebarPanel.js:
(WebInspector.NavigationSidebarPanel.prototype.matchTreeElementAgainstFilterFunctions):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182220 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago<rdar://problem/20365675> [iOS] Include Add to Reading List link action only where...
mitz@apple.com [Wed, 1 Apr 2015 04:06:07 +0000 (04:06 +0000)]
<rdar://problem/20365675> [iOS] Include Add to Reading List link action only where supported
https://bugs.webkit.org/show_bug.cgi?id=143289

Reviewed by Tim Horton.

* UIProcess/API/Cocoa/_WKElementAction.h: Excluded _WKElementActionTypeAddToReadingList from
the _WKElementActionType enum when Reading List is not supported.

* UIProcess/API/Cocoa/_WKElementAction.mm:
(+[_WKElementAction elementActionWithType:customTitle:]): Made Add to Reading List code
conditional on HAVE(SAFARI_SERVICES_FRAMEWORK).

* UIProcess/ios/WKActionSheetAssistant.mm:
(-[WKActionSheetAssistant showImageSheet]): Ditto.
(-[WKActionSheetAssistant showLinkSheet]): Ditto.

* config.h: Defined HAVE_SAFARI_SERVICES_FRAMEWORK.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182219 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoES6 class syntax should allow static setters and getters
rniwa@webkit.org [Wed, 1 Apr 2015 02:35:56 +0000 (02:35 +0000)]
ES6 class syntax should allow static setters and getters
https://bugs.webkit.org/show_bug.cgi?id=143180

Source/JavaScriptCore:

Reviewed by Filip Pizlo

Apparently I misread the spec when I initially implemented parseClass.
ES6 class syntax allows static getters and setters so just allow that.

* parser/Parser.cpp:
(JSC::Parser<LexerType>::parseClass):

LayoutTests:

Reviewed by Filip Pizlo.

Added test cases.

* js/class-syntax-declaration-expected.txt:
* js/class-syntax-expression-expected.txt:
* js/script-tests/class-syntax-declaration.js:
* js/script-tests/class-syntax-expression.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182218 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoTextIndicator for <span> inside an <a> only highlights the <span>, should highlight...
timothy_horton@apple.com [Wed, 1 Apr 2015 02:05:35 +0000 (02:05 +0000)]
TextIndicator for <span> inside an <a> only highlights the <span>, should highlight the whole <a>
https://bugs.webkit.org/show_bug.cgi?id=143287
<rdar://problem/20318342>

Reviewed by Beth Dakin.

* WebView/WebImmediateActionController.mm:
(-[WebImmediateActionController _defaultAnimationController]):
Instead of just highlighting the hit element, highlight the whole
element that the link URL comes from.

* WebProcess/WebPage/mac/WebPageMac.mm:
(WebKit::WebPage::performActionMenuHitTestAtLocation):
Instead of just highlighting the hit element, highlight the whole
element that the link URL comes from.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182217 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWKView should stop using endGestureWithEvent
timothy_horton@apple.com [Wed, 1 Apr 2015 01:15:58 +0000 (01:15 +0000)]
WKView should stop using endGestureWithEvent
https://bugs.webkit.org/show_bug.cgi?id=143285
<rdar://problem/20357007>

Reviewed by Simon Fraser.

* UIProcess/API/mac/WKView.mm:
(-[WKView magnifyWithEvent:]):
(-[WKView endGestureWithEvent:]): Deleted.
* UIProcess/mac/ViewGestureController.h:
* UIProcess/mac/ViewGestureControllerMac.mm:
(WebKit::ViewGestureController::endMagnificationGesture):
(WebKit::ViewGestureController::endActiveGesture): Deleted.
While there is no visible symptom, it is recommended that we check
the phase of normal magnification events instead of using the
special gesture end events.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182216 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMake it possible to compute a region for elements on the page that have wheel event...
simon.fraser@apple.com [Tue, 31 Mar 2015 23:42:31 +0000 (23:42 +0000)]
Make it possible to compute a region for elements on the page that have wheel event handlers
https://bugs.webkit.org/show_bug.cgi?id=142807

Reviewed by David Hyatt.

Source/WebCore:

Make it possible to compute a region that covers the elements on the page that have
a wheel event handler. This region may overestimate, but must not underestimate the area.

Elements with wheel handlers are registered on the document, and when a document gains
its first wheel handler, it registers the ownerElement in the parent document. Thus, on
the main frame, the region encompasses elements with handlers, and iframes whose subdocuments
have handlers.

Element gains some functions that allow it to return a rect which is the bounds of the element's
renderer, and renders for its descendant elements, which is the size or larger than the event
handling region for that element. Some configurations (e.g. position:fixed) require special
treatment.

Document::absoluteRegionForEventTargets() can then iterate over all elements in the given set,
and build a Region for those (short-circuiting if the document itself has a handler).

The new code is exercised for the debug MouseWheelRegionOverlay, and also added to the
non-fast scrollable region, used by threaded scrolling.

Tests: platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/handlers-in-iframes.html
       platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-fixed-child.html
       platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-in-columns.html
       platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-in-region.html
       platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-inside-fixed.html
       platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-on-document.html
       platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-on-fixed.html
       platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-region-basic.html

* dom/ContainerNode.h:
(WebCore::ContainerNode::absoluteEventHandlerBounds): Base class implementation; returns empty rect.
* dom/Document.cpp:
(WebCore::Document::prepareForDestruction): Unregister from the parent document.
(WebCore::Document::didAddWheelEventHandler): Add to the wheel handler set. Tell the parent
document if we are adding wheel handlers for the first time. Eventually the wheelEventHandlerCountChanged()
code will be removed, now we have the set. Tell debug overlays that we changed.
(WebCore::Document::didRemoveWheelEventHandler): Remove from the set, and unregister with the parent
document if necessary. Tell debug overlays that we changed.
(WebCore::Document::didAddTouchEventHandler): Minor cleanup.
(WebCore::Document::didRemoveTouchEventHandler): Ditto.
(WebCore::Document::didRemoveEventTargetNode): Remove from wheel targets.
(WebCore::Document::absoluteEventHandlerBounds): Implementation of the ContainerNode
function, just return the document bounds.
(WebCore::Document::absoluteRegionForEventTargets): Iterate over the given event targets,
and call absoluteEventHandlerBounds() on each.
* dom/Document.h:
(WebCore::Document::wheelEventTargets):
* dom/Element.cpp:
(WebCore::layoutOverflowRectContainsAllDescendants): Return true if we can determine that the
layoutOverflow bounds of the given renderer contains the renderers of all descendant elements.
(WebCore::Element::eventBounds): Get the "event handling" bounds of this element (excluding
descendants), and while doing so, compute whether any descendants are position:fixed, and whether
these bounds are known to include descendants.
(WebCore::Element::eventBoundsOfElementAndDescendants): Recursive function that short-circuits
if it can determine that descendants are enclosed by the bounds.
(WebCore::Element::absoluteEventHandlerBounds):
* dom/Element.h:
* page/DebugPageOverlays.cpp:
(WebCore::MouseWheelRegionOverlay::updateRegion): Ask the document for the region of wheel
event handlers.
* page/scrolling/ScrollingCoordinator.cpp:
(WebCore::ScrollingCoordinator::computeNonFastScrollableRegion): Add the wheel handler region
to the non-fast scrollable region.
* rendering/RenderLayer.cpp:
(WebCore::RenderLayer::scrollTo): Overflow scrolling needs to dirty event regions.
* rendering/RenderObject.cpp:
(WebCore::RenderObject::absoluteBoundingBoxRect): Need to pass the "wasFixed" flag out
of this function.
(WebCore::RenderObject::localToAbsolute): Ditto.
* rendering/RenderObject.h:

LayoutTests:

Tests that dump the non-fast-scrollable region, for various rendering configurations.

* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/handlers-in-iframes-expected.txt: Added.
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/handlers-in-iframes.html: Added.
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/resources/gain-wheel-handler.html: Added.
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/resources/lose-wheel-handler.html: Added.
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/resources/wheel-handler-region-helper.js: Added.
(rectsAsString):
(dumpRegion):
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-fixed-child-expected.txt: Added.
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-fixed-child.html: Added.
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-in-columns.html: Added.
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-in-region.html: Added.
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-inside-fixed-expected.txt: Added.
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-inside-fixed.html: Added.
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-on-document-expected.txt: Added.
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-on-document.html: Added.
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-on-fixed-expected.txt: Added.
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-on-fixed.html: Added.
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-region-basic-expected.txt: Added.
* platform/mac-wk2/tiled-drawing/scrolling/non-fast-region/wheel-handler-region-basic.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182215 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSwipes occur in the wrong direction in RTL
timothy_horton@apple.com [Tue, 31 Mar 2015 23:38:49 +0000 (23:38 +0000)]
Swipes occur in the wrong direction in RTL
https://bugs.webkit.org/show_bug.cgi?id=143257
<rdar://problem/20300862>

Reviewed by Simon Fraser.

* UIProcess/ios/ViewGestureControllerIOS.mm:
(-[WKSwipeTransitionController gestureRecognizerForInteractiveTransition:WithTarget:action:]):
When choosing which edge to attach our gesture recognizers to,
flip them in RTL.

(-[WKSwipeTransitionController directionForTransition:]):
(WebKit::ViewGestureController::beginSwipeGesture):
(WebKit::ViewGestureController::canSwipeInDirection):
* UIProcess/mac/ViewGestureController.h:
* UIProcess/mac/ViewGestureControllerMac.mm:
(WebKit::ViewGestureController::scrollEventCanBecomeSwipe):
(WebKit::ViewGestureController::trackSwipeGesture):
(WebKit::ViewGestureController::determineLayerAdjacentToSnapshotForParent):
(WebKit::ViewGestureController::beginSwipeGesture):
(WebKit::ViewGestureController::handleSwipeGesture):
Adjust SwipeDirection to be in logical back-forward list direction
instead of physical edge direction.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182214 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoPutClosureVar CSE def() rule has a wrong base
fpizlo@apple.com [Tue, 31 Mar 2015 23:15:02 +0000 (23:15 +0000)]
PutClosureVar CSE def() rule has a wrong base
https://bugs.webkit.org/show_bug.cgi?id=143280

Reviewed by Michael Saboff.

I think that this code was incorrect in a benign way, since the base of a
PutClosureVar is not a JS-visible object. But it was preventing some optimizations.

* dfg/DFGClobberize.h:
(JSC::DFG::clobberize):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182213 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix bindings tests results after <http://trac.webkit.org/changeset/182205>.
ap@apple.com [Tue, 31 Mar 2015 23:07:51 +0000 (23:07 +0000)]
Fix bindings tests results after <trac.webkit.org/changeset/182205>.

* bindings/scripts/test/JS/JSTestObj.cpp:
(WebCore::setJSTestObjTestSubObjEnabledBySettingConstructor):
(WebCore::setJSTestObjConditionalAttr4Constructor):
(WebCore::setJSTestObjConditionalAttr5Constructor):
(WebCore::setJSTestObjConditionalAttr6Constructor):
(WebCore::setJSTestObjReplaceableAttribute):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182212 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoContent Extensions: split the state machines to minimize prefix states
benjamin@webkit.org [Tue, 31 Mar 2015 22:59:21 +0000 (22:59 +0000)]
Content Extensions: split the state machines to minimize prefix states
https://bugs.webkit.org/show_bug.cgi?id=143253

Patch by Benjamin Poulain <bpoulain@apple.com> on 2015-03-31
Reviewed by Alex Christensen.

Source/WebCore:

With this patch, we refine how the URLs are partitioned to make
multiple small machines instead of a big one.

Here, considering each pattern as a language, I took the simplest
partitioning: take the union minimizing the states
that differentiate languages.

The idea is the following: every time a subgraph can be skipped
or repeated, it is an opportunity for a merged language to start
with a distinguishable state.
To avoid accumulating those differences, we only merge languages
that share the exact same "variable" subgraph as a prefix.

A side effect is that we could create as many machines as we have patterns.
We'll have to address that in a follow-up by merging small machines
until they reach a certain size.
In the common cases, patterns look very much alike so that is not a problem
at the moment.

* WebCore.xcodeproj/project.pbxproj:
* contentextensions/CombinedURLFilters.cpp: Added.
I moved the prefix tree and graph construction out of URLFilterParser for clarity.

CombinedURLFilters maintain the term prefix tree. Instead of generating on the fly
as we build the prefix tree, we first generate the entire tree of terms, then generate
NFAs for each branch without a variable prefix.

(WebCore::ContentExtensions::CombinedURLFilters::CombinedURLFilters):
(WebCore::ContentExtensions::CombinedURLFilters::~CombinedURLFilters):

(WebCore::ContentExtensions::CombinedURLFilters::addPattern):
Note here that we keep track of which subtree contains a any variable length term.
Instead of anotating the edges, the destination vertex is anotated with the information
from the edge.

The flag "inVariableLengthPrefix" is there to simplify traversal.

(WebCore::ContentExtensions::generateNFAForSubtree):
(WebCore::ContentExtensions::CombinedURLFilters::createNFAs):
Since we are traversing a tree generated by string, using the stack for recursion
is out of the question.

Fortunately, the entire state is very simple when going depth first: we only need
the active vertex and edge, and traverse the edges in order.

To do keep track of the edges, I used the HashTable iterator on the current term.
Since no vertex is modified for generation, storing the iterators is safe.

* contentextensions/CombinedURLFilters.h:
* contentextensions/ContentExtensionCompiler.cpp:
(WebCore::ContentExtensions::compileRuleList):
* contentextensions/NFA.cpp:
(WebCore::ContentExtensions::NFA::setActions):
(WebCore::ContentExtensions::NFA::debugPrintDot):
(WebCore::ContentExtensions::NFA::setFinal): Deleted.
(WebCore::ContentExtensions::NFA::addRuleId): Deleted.
* contentextensions/NFA.h:
* contentextensions/NFAGraphBuilder.cpp: Added.
* contentextensions/NFAGraphBuilder.h: Added.
* contentextensions/NFANode.h:
* contentextensions/Term.h: Added.
Move the Term to its own file since it is now shared between the parser
and the CombinedURLFilters.

(WebCore::ContentExtensions::Term::CharacterSet::operator==):
(WebCore::ContentExtensions::Term::CharacterSet::hash):
(WebCore::ContentExtensions::Term::Group::operator==):
(WebCore::ContentExtensions::Term::Group::hash):
(WebCore::ContentExtensions::TermHash::hash):
(WebCore::ContentExtensions::TermHash::equal):
(WebCore::ContentExtensions::Term::Term):
(WebCore::ContentExtensions::Term::~Term):
(WebCore::ContentExtensions::Term::isValid):
(WebCore::ContentExtensions::Term::addCharacter):
(WebCore::ContentExtensions::Term::extendGroupSubpattern):
(WebCore::ContentExtensions::Term::quantify):
(WebCore::ContentExtensions::Term::Term::generateGraph):
(WebCore::ContentExtensions::Term::isEndOfLineAssertion):
(WebCore::ContentExtensions::Term::matchesAtLeastOneCharacter):
(WebCore::ContentExtensions::Term::isKnownToMatchAnyString):
(WebCore::ContentExtensions::Term::hasFixedLength):
(WebCore::ContentExtensions::Term::operator=):
(WebCore::ContentExtensions::Term::operator==):
(WebCore::ContentExtensions::Term::hash):
(WebCore::ContentExtensions::Term::isEmptyValue):
(WebCore::ContentExtensions::Term::isDeletedValue):
(WebCore::ContentExtensions::Term::isUniversalTransition):
(WebCore::ContentExtensions::Term::generateSubgraphForAtom):
(WebCore::ContentExtensions::Term::destroy):
* contentextensions/URLFilterParser.cpp:
(WebCore::ContentExtensions::PatternParser::PatternParser):
(WebCore::ContentExtensions::PatternParser::finalize):
(WebCore::ContentExtensions::PatternParser::assertionBOL):
(WebCore::ContentExtensions::URLFilterParser::URLFilterParser):
(WebCore::ContentExtensions::URLFilterParser::addPattern):
(WebCore::ContentExtensions::Term::Term): Deleted.
(WebCore::ContentExtensions::Term::~Term): Deleted.
(WebCore::ContentExtensions::Term::isValid): Deleted.
(WebCore::ContentExtensions::Term::addCharacter): Deleted.
(WebCore::ContentExtensions::Term::extendGroupSubpattern): Deleted.
(WebCore::ContentExtensions::Term::quantify): Deleted.
(WebCore::ContentExtensions::Term::generateGraph): Deleted.
(WebCore::ContentExtensions::Term::isEndOfLineAssertion): Deleted.
(WebCore::ContentExtensions::Term::matchesAtLeastOneCharacter): Deleted.
(WebCore::ContentExtensions::Term::isKnownToMatchAnyString): Deleted.
(WebCore::ContentExtensions::Term::operator=): Deleted.
(WebCore::ContentExtensions::Term::operator==): Deleted.
(WebCore::ContentExtensions::Term::hash): Deleted.
(WebCore::ContentExtensions::Term::isEmptyValue): Deleted.
(WebCore::ContentExtensions::Term::isDeletedValue): Deleted.
(WebCore::ContentExtensions::Term::isUniversalTransition): Deleted.
(WebCore::ContentExtensions::Term::generateSubgraphForAtom): Deleted.
(WebCore::ContentExtensions::Term::destroy): Deleted.
(WebCore::ContentExtensions::Term::CharacterSet::operator==): Deleted.
(WebCore::ContentExtensions::Term::CharacterSet::hash): Deleted.
(WebCore::ContentExtensions::Term::Group::operator==): Deleted.
(WebCore::ContentExtensions::Term::Group::hash): Deleted.
(WebCore::ContentExtensions::TermHash::hash): Deleted.
(WebCore::ContentExtensions::TermHash::equal): Deleted.
(WebCore::ContentExtensions::GraphBuilder::GraphBuilder): Deleted.
(WebCore::ContentExtensions::GraphBuilder::finalize): Deleted.
(WebCore::ContentExtensions::GraphBuilder::assertionBOL): Deleted.
* contentextensions/URLFilterParser.h:

Tools:

* TestWebKitAPI/Tests/WebCore/ContentExtensions.cpp:
(TestWebKitAPI::TEST_F):
(TestWebKitAPI::testPatternStatus):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182211 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r182200.
commit-queue@webkit.org [Tue, 31 Mar 2015 22:45:02 +0000 (22:45 +0000)]
Unreviewed, rolling out r182200.
https://bugs.webkit.org/show_bug.cgi?id=143279

Probably causing assertion extravaganza on bots. (Requested by
kling on #webkit).

Reverted changeset:

"Logically empty WeakBlocks should not pin down their
MarkedBlocks indefinitely."
https://bugs.webkit.org/show_bug.cgi?id=143210
http://trac.webkit.org/changeset/182200

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182209 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMove ExpansionBehaviorFlags and TextDirection to their own file
mmaxfield@apple.com [Tue, 31 Mar 2015 21:34:33 +0000 (21:34 +0000)]
Move ExpansionBehaviorFlags and TextDirection to their own file
https://bugs.webkit.org/show_bug.cgi?id=143273

Reviewed by Simon Fraser.

This is a mechanical change. It just moves code around.

This is getting ready for https://bugs.webkit.org/show_bug.cgi?id=142608

Source/WebCore:

No new tests because there is no behavior change.

* Modules/notifications/Notification.h:
* WebCore.vcxproj/WebCore.vcxproj:
* WebCore.vcxproj/WebCore.vcxproj.filters:
* WebCore.xcodeproj/project.pbxproj:
* css/CSSPrimitiveValueMappings.h:
* css/CSSProperty.h:
* dom/Position.h:
* editing/VisiblePosition.h:
* editing/VisibleUnits.h:
* editing/htmlediting.h:
* html/canvas/CanvasRenderingContext2D.cpp:
(WebCore::CanvasRenderingContext2D::drawTextInternal):
* html/canvas/CanvasRenderingContext2D.h:
* platform/LengthBox.h:
* platform/PopupMenuStyle.h:
* platform/graphics/FontCascade.h:
* platform/graphics/LayoutBoxExtent.h:
* platform/graphics/TextRun.h:
* platform/ios/SelectionRect.h:
* platform/text/BidiResolver.h:
* platform/text/StringWithDirection.h:
* platform/text/TextHelpers.h: Renamed from Source/WebCore/platform/text/TextDirection.h.
(WebCore::isLeftToRightDirection):
* platform/win/PopupMenuWin.cpp:
(WebCore::PopupMenuWin::paint):
* rendering/EllipsisBox.cpp:
(WebCore::EllipsisBox::paint):
(WebCore::EllipsisBox::selectionRect):
(WebCore::EllipsisBox::paintSelection):
* rendering/HitTestLocation.h:
* rendering/HitTestResult.h:
* rendering/InlineBox.h:
* rendering/InlineTextBox.h:
(WebCore::InlineTextBox::expansionBehavior):
* rendering/RenderBlock.cpp:
(WebCore::constructTextRunInternal):
(WebCore::RenderBlock::constructTextRun):
* rendering/RenderBlock.h:
* rendering/RenderFileUploadControl.cpp:
(WebCore::RenderFileUploadControl::paintObject):
(WebCore::RenderFileUploadControl::computeIntrinsicLogicalWidths):
* rendering/RenderListBox.cpp:
(WebCore::RenderListBox::updateFromElement):
(WebCore::RenderListBox::paintItemForeground):
* rendering/RenderListMarker.cpp:
(WebCore::RenderListMarker::computePreferredLogicalWidths):
(WebCore::RenderListMarker::getRelativeMarkerRect):
* rendering/RenderMenuList.cpp:
(RenderMenuList::updateOptionsWidth):
* rendering/RenderTextControl.cpp:
(WebCore::RenderTextControl::getAverageCharWidth):
* rendering/RenderThemeIOS.mm:
* rendering/style/RenderStyle.h:
* rendering/svg/SVGInlineTextBox.cpp:
(WebCore::SVGInlineTextBox::constructTextRun):
* rendering/svg/SVGTextMetrics.cpp:
(WebCore::SVGTextMetrics::constructTextRun):

Source/WebKit/ios:

* WebCoreSupport/WebFrameIOS.mm:
* WebCoreSupport/WebVisiblePosition.mm:

Source/WebKit2:

* Shared/WebPopupItem.h:
* UIProcess/API/C/efl/WKAPICastEfl.h:
* UIProcess/API/efl/EwkView.h:
* UIProcess/WebPopupMenuProxy.h:
* UIProcess/efl/WebUIPopupMenuClient.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182207 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRemove user project file left in tree by accident.
jonowells@apple.com [Tue, 31 Mar 2015 21:34:07 +0000 (21:34 +0000)]
Remove user project file left in tree by accident.

* WebInspectorUI.esproj/Project.espressostorage: Removed.
* WebInspectorUI.esproj/user.jonowells.espressostorage: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182206 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoClean up Identifier factories to clarify the meaning of StringImpl*
utatane.tea@gmail.com [Tue, 31 Mar 2015 21:25:14 +0000 (21:25 +0000)]
Clean up Identifier factories to clarify the meaning of StringImpl*
https://bugs.webkit.org/show_bug.cgi?id=143146

Reviewed by Filip Pizlo.

Source/JavaScriptCore:

In the a lot of places, `Identifier(VM*/ExecState*, StringImpl*)` constructor is used.
However, it's ambiguous because `StringImpl*` has 2 different meanings.
1) normal string, it is replacable with `WTFString` and
2) `uid`, which holds `isSymbol` information to represent Symbols.
So we dropped Identifier constructors for strings and instead, introduced 2 factory functions.
+ `Identifier::fromString(VM*/ExecState*, const String&)`.
Just construct Identifier from strings. The symbol-ness of StringImpl* is not kept.
+ `Identifier::fromUid(VM*/ExecState*, StringImpl*)`.
This function is used for 2) `uid`. So symbol-ness of `StringImpl*` is kept.

And to clean up `StringImpl` which is used as uid,
we introduce `StringKind` into `StringImpl`. There's 3 kinds
1. StringNormal (non-atomic, non-symbol)
2. StringAtomic (atomic, non-symbol)
3. StringSymbol (non-atomic, symbol)
They are mutually exclusive. And (atomic, symbol) case should not exist.

* API/JSCallbackObjectFunctions.h:
(JSC::JSCallbackObject<Parent>::getOwnNonIndexPropertyNames):
* API/JSObjectRef.cpp:
(JSObjectMakeFunction):
* API/OpaqueJSString.cpp:
(OpaqueJSString::identifier):
* bindings/ScriptFunctionCall.cpp:
(Deprecated::ScriptFunctionCall::call):
* builtins/BuiltinExecutables.cpp:
(JSC::BuiltinExecutables::createExecutableInternal):
* builtins/BuiltinNames.h:
(JSC::BuiltinNames::BuiltinNames):
* bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::BytecodeGenerator):
(JSC::BytecodeGenerator::emitThrowReferenceError):
(JSC::BytecodeGenerator::emitThrowTypeError):
(JSC::BytecodeGenerator::emitReadOnlyExceptionIfNeeded):
(JSC::BytecodeGenerator::emitEnumeration):
* dfg/DFGDesiredIdentifiers.cpp:
(JSC::DFG::DesiredIdentifiers::reallyAdd):
* inspector/JSInjectedScriptHost.cpp:
(Inspector::JSInjectedScriptHost::functionDetails):
(Inspector::constructInternalProperty):
(Inspector::JSInjectedScriptHost::weakMapEntries):
(Inspector::JSInjectedScriptHost::iteratorEntries):
* inspector/JSInjectedScriptHostPrototype.cpp:
(Inspector::JSInjectedScriptHostPrototype::finishCreation):
* inspector/JSJavaScriptCallFramePrototype.cpp:
* inspector/ScriptCallStackFactory.cpp:
(Inspector::extractSourceInformationFromException):
* jit/JITOperations.cpp:
* jsc.cpp:
(GlobalObject::finishCreation):
(GlobalObject::addFunction):
(GlobalObject::addConstructableFunction):
(functionRun):
(runWithScripts):
* llint/LLIntData.cpp:
(JSC::LLInt::Data::performAssertions):
* llint/LowLevelInterpreter.asm:
* parser/ASTBuilder.h:
(JSC::ASTBuilder::addVar):
* parser/Parser.cpp:
(JSC::Parser<LexerType>::parseInner):
(JSC::Parser<LexerType>::createBindingPattern):
* parser/ParserArena.h:
(JSC::IdentifierArena::makeIdentifier):
(JSC::IdentifierArena::makeIdentifierLCharFromUChar):
(JSC::IdentifierArena::makeNumericIdentifier):
* runtime/ArgumentsIteratorPrototype.cpp:
(JSC::ArgumentsIteratorPrototype::finishCreation):
* runtime/ArrayIteratorPrototype.cpp:
(JSC::ArrayIteratorPrototype::finishCreation):
* runtime/ArrayPrototype.cpp:
(JSC::ArrayPrototype::finishCreation):
(JSC::arrayProtoFuncPush):
* runtime/ClonedArguments.cpp:
(JSC::ClonedArguments::getOwnPropertySlot):
* runtime/CommonIdentifiers.cpp:
(JSC::CommonIdentifiers::CommonIdentifiers):
* runtime/CommonIdentifiers.h:
* runtime/Error.cpp:
(JSC::addErrorInfo):
(JSC::hasErrorInfo):
* runtime/ExceptionHelpers.cpp:
(JSC::createUndefinedVariableError):
* runtime/GenericArgumentsInlines.h:
(JSC::GenericArguments<Type>::getOwnPropertySlot):
* runtime/Identifier.h:
(JSC::Identifier::isSymbol):
(JSC::Identifier::Identifier):
(JSC::Identifier::from): Deleted.
* runtime/IdentifierInlines.h:
(JSC::Identifier::Identifier):
(JSC::Identifier::fromUid):
(JSC::Identifier::fromString):
* runtime/JSCJSValue.cpp:
(JSC::JSValue::dumpInContextAssumingStructure):
* runtime/JSCJSValueInlines.h:
(JSC::JSValue::toPropertyKey):
* runtime/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init):
* runtime/JSLexicalEnvironment.cpp:
(JSC::JSLexicalEnvironment::getOwnNonIndexPropertyNames):
* runtime/JSObject.cpp:
(JSC::getClassPropertyNames):
(JSC::JSObject::reifyStaticFunctionsForDelete):
* runtime/JSObject.h:
(JSC::makeIdentifier):
* runtime/JSPromiseConstructor.cpp:
(JSC::JSPromiseConstructorFuncRace):
(JSC::JSPromiseConstructorFuncAll):
* runtime/JSString.h:
(JSC::JSString::toIdentifier):
* runtime/JSSymbolTableObject.cpp:
(JSC::JSSymbolTableObject::getOwnNonIndexPropertyNames):
* runtime/LiteralParser.cpp:
(JSC::LiteralParser<CharType>::tryJSONPParse):
(JSC::LiteralParser<CharType>::makeIdentifier):
* runtime/Lookup.h:
(JSC::reifyStaticProperties):
* runtime/MapConstructor.cpp:
(JSC::constructMap):
* runtime/MapIteratorPrototype.cpp:
(JSC::MapIteratorPrototype::finishCreation):
* runtime/MapPrototype.cpp:
(JSC::MapPrototype::finishCreation):
* runtime/MathObject.cpp:
(JSC::MathObject::finishCreation):
* runtime/NumberConstructor.cpp:
(JSC::NumberConstructor::finishCreation):
* runtime/ObjectConstructor.cpp:
(JSC::ObjectConstructor::finishCreation):
* runtime/PrivateName.h:
(JSC::PrivateName::PrivateName):
* runtime/PropertyMapHashTable.h:
(JSC::PropertyTable::find):
(JSC::PropertyTable::get):
* runtime/PropertyName.h:
(JSC::PropertyName::PropertyName):
(JSC::PropertyName::publicName):
(JSC::PropertyName::asIndex):
* runtime/PropertyNameArray.cpp:
(JSC::PropertyNameArray::add):
* runtime/PropertyNameArray.h:
(JSC::PropertyNameArray::addKnownUnique):
* runtime/RegExpConstructor.cpp:
(JSC::RegExpConstructor::finishCreation):
* runtime/SetConstructor.cpp:
(JSC::constructSet):
* runtime/SetIteratorPrototype.cpp:
(JSC::SetIteratorPrototype::finishCreation):
* runtime/SetPrototype.cpp:
(JSC::SetPrototype::finishCreation):
* runtime/StringIteratorPrototype.cpp:
(JSC::StringIteratorPrototype::finishCreation):
* runtime/StringPrototype.cpp:
(JSC::StringPrototype::finishCreation):
* runtime/Structure.cpp:
(JSC::Structure::getPropertyNamesFromStructure):
* runtime/SymbolConstructor.cpp:
* runtime/VM.cpp:
(JSC::VM::throwException):
* runtime/WeakMapConstructor.cpp:
(JSC::constructWeakMap):

Source/WebCore:

Just change Identifier creations.

1. If the code creates Identifier from StringImpl*
which is treated as symbol or string(unique id), use Identifier::fromUid.
2. If the code creates Identifier from string, use Identifier::fromString.

* Modules/plugins/QuickTimePluginReplacement.mm:
(WebCore::QuickTimePluginReplacement::ensureReplacementScriptInjected):
(WebCore::QuickTimePluginReplacement::installReplacement):
* bindings/js/IDBBindingUtilities.cpp:
(WebCore::get):
(WebCore::set):
* bindings/js/JSCSSStyleDeclarationCustom.cpp:
(WebCore::JSCSSStyleDeclaration::getOwnPropertyNames):
* bindings/js/JSCallbackData.cpp:
(WebCore::JSCallbackData::invokeCallback):
* bindings/js/JSCommandLineAPIHostCustom.cpp:
(WebCore::getJSListenerFunctions):
(WebCore::JSCommandLineAPIHost::getEventListeners):
* bindings/js/JSCryptoAlgorithmBuilder.cpp:
(WebCore::JSCryptoAlgorithmBuilder::add):
* bindings/js/JSCryptoAlgorithmDictionary.cpp:
(WebCore::getProperty):
(WebCore::getHashAlgorithm):
* bindings/js/JSCryptoKeySerializationJWK.cpp:
(WebCore::getJSArrayFromJSON):
(WebCore::getStringFromJSON):
(WebCore::getBooleanFromJSON):
(WebCore::addToJSON):
(WebCore::buildJSONForRSAComponents):
(WebCore::addBoolToJSON):
(WebCore::addUsagesToJSON):
* bindings/js/JSCustomXPathNSResolver.cpp:
(WebCore::JSCustomXPathNSResolver::lookupNamespaceURI):
* bindings/js/JSDOMStringMapCustom.cpp:
(WebCore::JSDOMStringMap::getOwnPropertyNames):
* bindings/js/JSDOMWindowCustom.cpp:
(WebCore::JSDOMWindow::defineOwnProperty):
(WebCore::JSDOMWindow::setLocation):
(WebCore::DialogHandler::dialogCreated):
(WebCore::DialogHandler::returnValue):
* bindings/js/JSDeviceMotionEventCustom.cpp:
(WebCore::readAccelerationArgument):
(WebCore::readRotationRateArgument):
(WebCore::createAccelerationObject):
(WebCore::createRotationRateObject):
* bindings/js/JSDictionary.cpp:
(WebCore::JSDictionary::tryGetProperty):
* bindings/js/JSEventListener.cpp:
(WebCore::JSEventListener::handleEvent):
* bindings/js/JSHTMLAllCollectionCustom.cpp:
(WebCore::callHTMLAllCollection):
(WebCore::JSHTMLAllCollection::item):
(WebCore::JSHTMLAllCollection::namedItem):
* bindings/js/JSHTMLDocumentCustom.cpp:
(WebCore::JSHTMLDocument::all):
(WebCore::JSHTMLDocument::setAll):
(WebCore::JSHTMLDocument::open):
* bindings/js/JSHTMLFormControlsCollectionCustom.cpp:
(WebCore::JSHTMLFormControlsCollection::namedItem):
* bindings/js/JSIDBDatabaseCustom.cpp:
(WebCore::JSIDBDatabase::createObjectStore):
* bindings/js/JSIDBObjectStoreCustom.cpp:
(WebCore::JSIDBObjectStore::createIndex):
* bindings/js/JSImageDataCustom.cpp:
(WebCore::toJS):
* bindings/js/JSInspectorFrontendHostCustom.cpp:
(WebCore::populateContextMenuItems):
* bindings/js/JSLazyEventListener.cpp:
(WebCore::JSLazyEventListener::initializeJSFunction):
* bindings/js/JSNodeFilterCondition.cpp:
(WebCore::JSNodeFilterCondition::acceptNode):
* bindings/js/JSSQLResultSetRowListCustom.cpp:
(WebCore::JSSQLResultSetRowList::item):
* bindings/js/JSStorageCustom.cpp:
(WebCore::JSStorage::getOwnPropertyNames):
* bindings/js/ReadableStreamJSSource.cpp:
(WebCore::setInternalSlotToObject):
(WebCore::getInternalSlotFromObject):
* bindings/js/ScriptGlobalObject.cpp:
(WebCore::ScriptGlobalObject::set):
(WebCore::ScriptGlobalObject::get):
(WebCore::ScriptGlobalObject::remove):
* bindings/js/SerializedScriptValue.cpp:
(WebCore::CloneSerializer::CloneSerializer):
(WebCore::CloneSerializer::write):
(WebCore::CloneDeserializer::deserialize):
* bindings/objc/WebScriptObject.mm:
(-[WebScriptObject callWebScriptMethod:withArguments:]):
(-[WebScriptObject setValue:forKey:]):
(-[WebScriptObject valueForKey:]):
(-[WebScriptObject removeWebScriptKey:]):
(-[WebScriptObject hasWebScriptKey:]):
* bindings/scripts/CodeGeneratorJS.pm:
(GenerateImplementation):
* bindings/scripts/test/JS/JSFloat64Array.cpp:
(WebCore::JSFloat64Array::getOwnPropertyNames):
* bindings/scripts/test/JS/JSTestEventTarget.cpp:
(WebCore::JSTestEventTarget::getOwnPropertyNames):
* bindings/scripts/test/JS/JSTestObj.cpp:
(WebCore::setJSTestObjTestSubObjEnabledBySettingConstructor):
(WebCore::setJSTestObjConditionalAttr4Constructor):
(WebCore::setJSTestObjConditionalAttr5Constructor):
(WebCore::setJSTestObjConditionalAttr6Constructor):
(WebCore::setJSTestObjReplaceableAttribute):
* bridge/c/c_utility.cpp:
(JSC::Bindings::identifierFromNPIdentifier):
* bridge/objc/objc_runtime.mm:
(JSC::Bindings::ObjcFallbackObjectImp::defaultValue):
* bridge/testbindings.cpp:
(main):
* bridge/testbindings.mm:
(main):
* contentextensions/ContentExtensionParser.cpp:
(WebCore::ContentExtensions::loadTrigger):
(WebCore::ContentExtensions::loadAction):
* html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::parseAttribute):
* html/HTMLPlugInImageElement.cpp:
(WebCore::HTMLPlugInImageElement::didAddUserAgentShadowRoot):
* testing/js/WebCoreTestSupport.cpp:
(WebCoreTestSupport::injectInternalsObject):

Source/WebKit/mac:

Just change to Identifier::fromString.

* Plugins/Hosted/NetscapePluginHostProxy.mm:
(identifierFromIdentifierRep):
* Plugins/Hosted/ProxyInstance.mm:
(WebKit::ProxyInstance::getPropertyNames):

Source/WebKit2:

Just change to Identifier::fromString.

* WebProcess/Plugins/Netscape/JSNPObject.cpp:
(WebKit::JSNPObject::getOwnPropertyNames):
* WebProcess/Plugins/Netscape/NPJSObject.cpp:
(WebKit::identifierFromIdentifierRep):

Source/WTF:

Introduce StringKind into `StringImpl`. There's 3 kinds
1. StringNormal (non-atomic, non-symbol)
2. StringAtomic (atomic, non-symbol)
3. StringSymbol (non-atomic, symbol)
They are mutually exclusive.

* wtf/text/AtomicString.cpp:
(WTF::AtomicString::addSlowCase):
(WTF::AtomicString::findSlowCase):
When registering a string into AtomicStringTable,
it should not be a symbol string
because symbol and atomic types are mutually exclusive.
When a symbol string comes, we extract an owner string
from a symbol string by using `StringImpl::extractFoldedStringInSymbol()`.
It always succeeds because a symbol (non empty) string
is always BufferSubstring and has an owner string.
Empty symbol string doesn't have an owner string.
This case is filtered by `!string.length()` guard.
* wtf/text/AtomicString.h:
(WTF::AtomicString::add):
(WTF::AtomicString::addWithStringTableProvider):
* wtf/text/StringImpl.cpp:
(WTF::StringImpl::~StringImpl):
(WTF::StringImpl::createSymbol):
(WTF::StringImpl::createUnique): Deleted.
* wtf/text/StringImpl.h:
(WTF::StringImpl::StringImpl):
(WTF::StringImpl::createSymbolEmpty):
(WTF::StringImpl::flagIsAtomic):
(WTF::StringImpl::flagIsSymbol):
(WTF::StringImpl::maskStringKind):
(WTF::StringImpl::stringKind):
(WTF::StringImpl::isSymbol):
(WTF::StringImpl::isAtomic):
(WTF::StringImpl::setIsAtomic):
(WTF::StringImpl::extractFoldedStringInSymbol):
(WTF::StringImpl::createUniqueEmpty): Deleted.
(WTF::StringImpl::flagIsUnique): Deleted.
(WTF::StringImpl::isUnique): Deleted.
* wtf/text/StringStatics.cpp:
(WTF::StringImpl::hashAndFlagsForSymbol):
(WTF::StringImpl::hashAndFlagsForUnique): Deleted.

Tools:

Simple API tests for StringImpl are added.

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

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182205 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Mac] Songza.com fails to play; QTKit path reports it is always "seeking".
jer.noble@apple.com [Tue, 31 Mar 2015 21:18:30 +0000 (21:18 +0000)]
[Mac] Songza.com fails to play; QTKit path reports it is always "seeking".
https://bugs.webkit.org/show_bug.cgi?id=143274

Rubber-stamped by Eric Carlson.

MediaTime::invalidTime() is always greater than any valid time. So when
checking to see if MediaPlayerPrivateQTKit is seeking, first check if the
m_seekTo time is valid before checking whether it's > 0.

* platform/graphics/mac/MediaPlayerPrivateQTKit.mm:
(WebCore::MediaPlayerPrivateQTKit::seeking):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182204 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWebKit briefly shows wrong webpage after swiping back (gigaom.com, or any site on...
timothy_horton@apple.com [Tue, 31 Mar 2015 20:53:32 +0000 (20:53 +0000)]
WebKit briefly shows wrong webpage after swiping back (gigaom.com, or any site on a slow network)
https://bugs.webkit.org/show_bug.cgi?id=143256
<rdar://problem/19458648>

Reviewed by Dan Bernstein.

Instead of allowing a flash of the previous page when the swipe snapshot
timeouts fire (removing the snapshot before the new page is loaded), show
the snapshotted background color.

This fixes the problem on iOS, where UI-side compositing makes it easy to fix,
but not yet on OS X.

* UIProcess/API/Cocoa/WKWebView.mm:
(baseScrollViewBackgroundColor):
(scrollViewBackgroundColor):
If the ViewGestureController returns a valid background color, use that
instead of the page's background color.

* UIProcess/API/Cocoa/WKWebViewInternal.h:
Expose _updateScrollViewBackground.

* UIProcess/PageClient.h:
* UIProcess/ios/PageClientImplIOS.h:
* UIProcess/ios/PageClientImplIOS.mm:
(WebKit::PageClientImpl::didChangeBackgroundColor):
* UIProcess/mac/PageClientImpl.h:
* UIProcess/mac/PageClientImpl.mm:
(WebKit::PageClientImpl::didChangeBackgroundColor):
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::didChangeBackgroundColor):
* UIProcess/WebPageProxy.h:
Add and plumb didChangeBackgroundColor, which calls
_updateScrollViewBackground on iOS.

* UIProcess/ios/ViewGestureControllerIOS.mm:
(WebKit::ViewGestureController::endSwipeGesture):
Hide the content of the drawing area until the next commit. This way,
even if the snapshot is removed (say, because a timeout fired), we won't
ever show the old page content (but we will show background color).

Store the background color associated with the current snapshot.

Let WKWebView know that it needs to recompute the background color.

(WebKit::ViewGestureController::removeSwipeSnapshot):
Clear the background color so that the next time we commit, WKWebView
will get an invalid color from ViewGestureController and fall back
to the page's extended background color instead.

* UIProcess/mac/ViewGestureController.h:
(WebKit::ViewGestureController::backgroundColorForCurrentSnapshot):
* UIProcess/mac/ViewGestureControllerMac.mm:
(WebKit::ViewGestureController::beginSwipeGesture):
(WebKit::ViewGestureController::removeSwipeSnapshot):
Keep backgroundColorForCurrentSnapshot up to date on Mac too, even
though we don't use it yet.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182203 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[iOS] Rotating PDF in Safari scrolls to the wrong position
timothy_horton@apple.com [Tue, 31 Mar 2015 20:50:12 +0000 (20:50 +0000)]
[iOS] Rotating PDF in Safari scrolls to the wrong position
https://bugs.webkit.org/show_bug.cgi?id=143259
<rdar://problem/19872693>

Reviewed by Dan Bernstein.

* UIProcess/ios/WKPDFView.mm:
(-[WKPDFView web_setMinimumSize:]):
Maintan the relative top/left within the rescaled document.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182202 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoaccessibility/aria-hidden-hides-all-elements.html fails on Windows
dino@apple.com [Tue, 31 Mar 2015 20:37:03 +0000 (20:37 +0000)]
accessibility/aria-hidden-hides-all-elements.html fails on Windows
https://bugs.webkit.org/show_bug.cgi?id=143275

Mark as Pass/Fail.

* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182201 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoLogically empty WeakBlocks should not pin down their MarkedBlocks indefinitely.
akling@apple.com [Tue, 31 Mar 2015 20:01:29 +0000 (20:01 +0000)]
Logically empty WeakBlocks should not pin down their MarkedBlocks indefinitely.
<https://webkit.org/b/143210>

Reviewed by Geoffrey Garen.

Since a MarkedBlock cannot be destroyed until all the WeakBlocks pointing into it are gone,
we had a little problem where WeakBlocks with only null pointers would still keep their
MarkedBlock alive.

This patch fixes that by detaching WeakBlocks from their MarkedBlock once a sweep discovers
that the WeakBlock contains no pointers to live objects. Ownership of the WeakBlock is passed
to the Heap, which will sweep the list of these detached WeakBlocks as part of a full GC,
destroying them once they're fully dead.

This allows the garbage collector to reclaim the 64kB MarkedBlocks much sooner, and resolves
a mysterious issue where doing two full garbage collections back-to-back would free additional
memory in the second collection.

Management of detached WeakBlocks is implemented as a Vector<WeakBlock*> in Heap, along with
an index of the next block in that vector that needs to be swept. The IncrementalSweeper then
calls into Heap::sweepNextLogicallyEmptyWeakBlock() to sweep one block at a time.

* heap/Heap.h:
* heap/Heap.cpp:
(JSC::Heap::collectAllGarbage): Add a final pass where we sweep the logically empty WeakBlocks
owned by Heap, after everything else has been swept.

(JSC::Heap::notifyIncrementalSweeper): Set up an incremental sweep of logically empty WeakBlocks
after a full garbage collection ends. Note that we don't do this after Eden collections, since
they are unlikely to cause entire WeakBlocks to go empty.

(JSC::Heap::addLogicallyEmptyWeakBlock): Added. Interface for passing ownership of a WeakBlock
to the Heap when it's detached from a WeakSet.

(JSC::Heap::sweepAllLogicallyEmptyWeakBlocks): Helper for collectAllGarbage() that sweeps all
of the logically empty WeakBlocks owned by Heap.

(JSC::Heap::sweepNextLogicallyEmptyWeakBlock): Sweeps one logically empty WeakBlock if needed
and updates the next-logically-empty-weak-block-to-sweep index.

(JSC::Heap::lastChanceToFinalize): call sweepAllLogicallyEmptyWeakBlocks() here, since there
won't be another chance after this.

* heap/IncrementalSweeper.h:
(JSC::IncrementalSweeper::hasWork): Deleted.

* heap/IncrementalSweeper.cpp:
(JSC::IncrementalSweeper::fullSweep):
(JSC::IncrementalSweeper::doSweep):
(JSC::IncrementalSweeper::sweepNextBlock): Restructured IncrementalSweeper a bit to simplify
adding a new sweeping stage for the Heap's logically empty WeakBlocks. sweepNextBlock() is
changed to return a bool (true if there's more work to be done.)

* heap/WeakBlock.cpp:
(JSC::WeakBlock::sweep): This now figures out if the WeakBlock is logically empty, i.e doesn't
contain any pointers to live objects. The answer is stored in a new SweepResult member.

* heap/WeakBlock.h:
(JSC::WeakBlock::isLogicallyEmptyButNotFree): Added. Can be queried after a sweep to determine
if the WeakBlock could be detached from the MarkedBlock.

(JSC::WeakBlock::SweepResult::SweepResult): Deleted in favor of initializing member variables
when declaring them.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182200 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoREGRESSION: Searching commits can highlight wrong data points
rniwa@webkit.org [Tue, 31 Mar 2015 19:43:45 +0000 (19:43 +0000)]
REGRESSION: Searching commits can highlight wrong data points
https://bugs.webkit.org/show_bug.cgi?id=143272

Reviewed by Antti Koivisto.

The bug was caused by /api/commits returning commit times with millisecond precision whereas /api/runs
return commit times with only second precision. This resulted in the frontend code to match a commit
with the data point that included the next commit when the millisecond component of commit's timestamp
wasn't identically 0.

This discrepancy was caused by the fact PHP's strtotime only ignores milliseconds and /api/commits
was returning timestamp as string instead of parsing via Database::to_js_time as done in /api/runs
so miliseconds component was only preserved in /api/commits.

Fixed the bug by always using Database::to_js_time to return commit time. Also fixed to_js_time so that
it returns time in milisecond precision.

* public/api/commits.php:
(fetch_commits_between): Use Database::to_js_time for format commit times.
(format_commit): Ditto.
* public/include/db.php:
(Database::to_js_time): Parse and append millisecond component. Ignore sub-milliseconds for simplicity.
* public/v2/data.js:
(CommitLogs.fetchForTimeRange): The commit time is now an integer so don't call "replace" on it.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182199 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoeval("this.foo") causes a crash if this had not been initialized in a derived class...
rniwa@webkit.org [Tue, 31 Mar 2015 19:42:56 +0000 (19:42 +0000)]
eval("this.foo") causes a crash if this had not been initialized in a derived class's constructor
https://bugs.webkit.org/show_bug.cgi?id=142883

Reviewed by Filip Pizlo.

The crash was caused by eval inside the constructor of a derived class not checking TDZ.

Fixed the bug by adding a parser flag that forces the TDZ check to be always emitted when accessing "this"
in eval inside a derived class' constructor.

* bytecode/EvalCodeCache.h:
(JSC::EvalCodeCache::getSlow):
* bytecompiler/NodesCodegen.cpp:
(JSC::ThisNode::emitBytecode):
* debugger/DebuggerCallFrame.cpp:
(JSC::DebuggerCallFrame::evaluate):
* interpreter/Interpreter.cpp:
(JSC::eval):
* parser/ASTBuilder.h:
(JSC::ASTBuilder::thisExpr):
* parser/NodeConstructors.h:
(JSC::ThisNode::ThisNode):
* parser/Nodes.h:
* parser/Parser.cpp:
(JSC::Parser<LexerType>::Parser):
(JSC::Parser<LexerType>::parsePrimaryExpression):
* parser/Parser.h:
(JSC::parse):
* parser/ParserModes.h:
* parser/SyntaxChecker.h:
(JSC::SyntaxChecker::thisExpr):
* runtime/CodeCache.cpp:
(JSC::CodeCache::getGlobalCodeBlock):
(JSC::CodeCache::getProgramCodeBlock):
(JSC::CodeCache::getEvalCodeBlock):
* runtime/CodeCache.h:
(JSC::SourceCodeKey::SourceCodeKey):
* runtime/Executable.cpp:
(JSC::EvalExecutable::create):
* runtime/Executable.h:
* runtime/JSGlobalObject.cpp:
(JSC::JSGlobalObject::createEvalCodeBlock):
* runtime/JSGlobalObject.h:
* runtime/JSGlobalObjectFunctions.cpp:
(JSC::globalFuncEval):
* tests/stress/class-syntax-no-tdz-in-eval.js: Added.
* tests/stress/class-syntax-tdz-in-eval.js: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182198 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoborder-image with 'fill' keyword does not fill the middle area unless the border...
commit-queue@webkit.org [Tue, 31 Mar 2015 19:32:12 +0000 (19:32 +0000)]
border-image with 'fill' keyword does not fill the middle area unless the border width is greater than zero.
https://bugs.webkit.org/show_bug.cgi?id=142650.

Patch by Said Abou-Hallawa <sabouhallawa@apple.com> on 2015-03-31
Reviewed by Simon Fraser.

Source/WebCore:

The decoration of a RenderBox, a RenderTable or an InlineFlowBox should be
drawn if its RenderStyle has a non-zero width border or the border-image
has the keyword fill.

Tests: fast/borders/border-image-fill-inline-no-border.html
       fast/borders/border-image-fill-no-border.html

* rendering/InlineFlowBox.cpp:
(WebCore::InlineFlowBox::paintBoxDecorations):
* rendering/RenderBox.cpp:
(WebCore::RenderBox::paintBoxDecorations):
* rendering/RenderBoxModelObject.cpp:
(WebCore::RenderBoxModelObject::hasBoxDecorationStyle):
* rendering/RenderTable.cpp:
(WebCore::RenderTable::paintBoxDecorations):
* rendering/style/BorderData.h:
(WebCore::BorderData::hasFill):
* rendering/style/RenderStyle.h:

LayoutTests:

Add tests to ensure the middle area of a RenderBox is going to be drawn
even if the border width is not greater than zero.

* fast/borders/border-image-fill-inline-no-border-expected.html: Added.
* fast/borders/border-image-fill-inline-no-border.html: Added.
* fast/borders/border-image-fill-no-border-expected.html: Added.
* fast/borders/border-image-fill-no-border.html: Added.
* fast/borders/resources/button-border-cropped.svg: Added.
* fast/borders/resources/button-border.svg: Added.
* fast/borders/resources/svg-100x100-intrinsic.svg: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182197 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Win] Unreviewed build fix.
achristensen@apple.com [Tue, 31 Mar 2015 19:25:32 +0000 (19:25 +0000)]
[Win] Unreviewed build fix.

* WebCore.vcxproj/WebCore.vcxproj:
Don't include inspector files twice in debug builds.
InspectorAllInOne.cpp is now used on debug builds.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182196 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[New Block-Inside-Inline Model] Anonymous inline-blocks should size as though they...
hyatt@apple.com [Tue, 31 Mar 2015 19:14:19 +0000 (19:14 +0000)]
[New Block-Inside-Inline Model] Anonymous inline-blocks should size as though they are block-level.
https://bugs.webkit.org/show_bug.cgi?id=143271

Reviewed by Dean Jackson.

Source/WebCore:

Added fast/block/inside-inlines/block-width-check.html.

* rendering/RenderBox.cpp:
(WebCore::RenderBox::sizesLogicalWidthToFitContent):
Return false when an anonymous inline-block is encountered, since we want to treat it like a regular
block instead.

LayoutTests:

* fast/block/inside-inlines/block-width-check-expected.html: Added.
* fast/block/inside-inlines/block-width-check.html: Added.
* fast/block/inside-inlines/new-model/block-width-check-expected.html: Added.
* fast/block/inside-inlines/new-model/block-width-check.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182195 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[WK2][iOS] Extend ProcessThrottler use to the NetworkProcess
cdumez@apple.com [Tue, 31 Mar 2015 19:03:40 +0000 (19:03 +0000)]
[WK2][iOS] Extend ProcessThrottler use to the NetworkProcess
https://bugs.webkit.org/show_bug.cgi?id=143249
<rdar://problem/20231539>

Reviewed by Gavin Barraclough.

Extend ProcessThrottler use to the NetworkProcess instead of only the
WebProcesses. This is needed so that the NetworkProcess gets notified
before suspension. We now use this notification to free-up memory
before suspending by calling the critical memory pressure handler, as
we already do for WebProcesses (rdar://problem/20231539). In the
future, this infrastructure will be used to delay suspension until the
pending SQL transactions are done as well (rdar://problem/20180799).

Previously, the NetworkProcessProxy would hold a foreground assertion
on behalf of the NetworkProcess but would never release that assertion.
In this patch, we change this so that WebProcessProxies hold assertions
on behalf of the NetworkProcess, and release those assertions before
suspending. As a result, the NetworkProcess gets suspended when all
WebProcesses are suspended.

* NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::processWillSuspend):
(WebKit::NetworkProcess::cancelProcessWillSuspend):
(WebKit::NetworkProcess::processDidResume):
* NetworkProcess/NetworkProcess.h:
* NetworkProcess/NetworkProcess.messages.in:
* UIProcess/Network/NetworkProcessProxy.cpp:
(WebKit::NetworkProcessProxy::NetworkProcessProxy):
(WebKit::NetworkProcessProxy::didFinishLaunching):
(WebKit::NetworkProcessProxy::sendProcessWillSuspend):
(WebKit::NetworkProcessProxy::sendCancelProcessWillSuspend):
(WebKit::NetworkProcessProxy::didCancelProcessSuspension):
(WebKit::NetworkProcessProxy::sendProcessDidResume):
(WebKit::NetworkProcessProxy::processReadyToSuspend):
* UIProcess/Network/NetworkProcessProxy.h:
(WebKit::NetworkProcessProxy::throttler):
* UIProcess/Network/NetworkProcessProxy.messages.in:
* UIProcess/ProcessThrottler.cpp:
(WebKit::ProcessThrottler::ProcessThrottler):
(WebKit::ProcessThrottler::didConnectToProcess):
(WebKit::ProcessThrottler::didConnnectToProcess): Deleted.
* UIProcess/ProcessThrottler.h:
* UIProcess/ProcessThrottlerClient.h: Added.
(WebKit::ProcessThrottlerClient::~ProcessThrottlerClient):
* UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::ensureNetworkProcess):
* UIProcess/WebProcessPool.h:
* UIProcess/WebProcessProxy.cpp:
(WebKit::WebProcessProxy::didFinishLaunching):
(WebKit::WebProcessProxy::initializeNetworkProcessActivityToken):
(WebKit::WebProcessProxy::sendProcessDidResume):
(WebKit::WebProcessProxy::processReadyToSuspend):
* UIProcess/WebProcessProxy.h:
* WebKit2.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182194 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r182186.
commit-queue@webkit.org [Tue, 31 Mar 2015 19:01:59 +0000 (19:01 +0000)]
Unreviewed, rolling out r182186.
https://bugs.webkit.org/show_bug.cgi?id=143270

it crashes all the WebGL tests on the Debug bots (Requested by
dino on #webkit).

Reverted changeset:

"Web Inspector: add 2D/WebGL canvas instrumentation
infrastructure"
https://bugs.webkit.org/show_bug.cgi?id=137278
http://trac.webkit.org/changeset/182186

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182193 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoCrash in CGContextShowGlyphsWithAdvances when passing kCGFontIndexInvalid
mmaxfield@apple.com [Tue, 31 Mar 2015 18:53:25 +0000 (18:53 +0000)]
Crash in CGContextShowGlyphsWithAdvances when passing kCGFontIndexInvalid
https://bugs.webkit.org/show_bug.cgi?id=143114

This is a workaround for <rdar://problem/20230073>. Please remove when it is no longer necessary.

Reviewed by Alexey Proskuryakov.

Covered by:
compositing/regions/floated-region-with-transformed-child.html
compositing/regions/floated-region-with-transformed-child-expected.html
fast/regions/counters/extract-ordered-lists-in-regions-explicit-counters-005.html
fast/regions/counters/extract-ordered-lists-in-regions-explicit-counters-005-expected.html
fast/regions/overflow/overflow-content-transform-rotate.html
fast/regions/overflow/overflow-content-transform-rotate-expected.html

* platform/graphics/GlyphBuffer.h:
(WebCore::GlyphBuffer::shrink): Performing shaping may remove glyphs, so we need to shrink the GlyphBuffer.
* platform/graphics/WidthIterator.cpp:
(WebCore::applyFontTransforms): Filter out kCGFontIndexInvalid.
(WebCore::WidthIterator::advanceInternal): Moved code into applyFontTransforms, and trigger the
shrink of the GlyphBuffer.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182192 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoREGRESSION (r173484): Reducing content of scrollable region does not reset scroll
bdakin@apple.com [Tue, 31 Mar 2015 18:26:16 +0000 (18:26 +0000)]
REGRESSION (r173484): Reducing content of scrollable region does not reset scroll
position
https://bugs.webkit.org/show_bug.cgi?id=138525
-and corresponding-
rdar://problem/18166043

Reviewed by Simon Fraser.

Source/WebCore:

The change that caused this regression was correct. That change does not allow
RenderLayer to update scroll position after a layout if a rubber-band is currently
happening. The change caused this regression because all of the member variables
in ScrollController that attempt to keep track of the current state of the scroll
gesture (m_inScrollGesture, m_momentumScrollInProgress, and
m_snapRubberbandTimerIsActive) all indicated that a momentum scroll gesture was
still in action for this div even though it very much is not when the bug happens.
Those variables were never properly re-set because the
PlatformWheelEventPhaseEnded events never got dispatched to the ScrollController,
which brought the investigation back to Element.

We must still dispatch events that have zero delta so that the default event
handlers can handle them, but we should stopPropagation() so that these events are
not sent to the DOM. Websites will break if they get wheel events with no delta.
* dom/Element.cpp:
(WebCore::Element::dispatchWheelEvent):

LayoutTests:

* platform/mac-wk2/tiled-drawing/scrolling/overflow-scroll-reduced-content-expected.txt: Added.
* platform/mac-wk2/tiled-drawing/scrolling/overflow-scroll-reduced-content.html: Added.
* platform/mac-wk2/tiled-drawing/scrolling/overflow-scroll-zero-delta-wheel-events-expected.txt: Added.
* platform/mac-wk2/tiled-drawing/scrolling/overflow-scroll-zero-delta-wheel-events.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182191 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Win] Unreviewed debug build fix after r182186.
achristensen@apple.com [Tue, 31 Mar 2015 18:22:32 +0000 (18:22 +0000)]
[Win] Unreviewed debug build fix after r182186.

* WebCore.vcxproj/WebCore.vcxproj:
Include InspectorAllInOne.cpp in debug builds.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182190 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[ES6] Object type restrictions on a first parameter of several Object.* functions...
utatane.tea@gmail.com [Tue, 31 Mar 2015 17:44:37 +0000 (17:44 +0000)]
[ES6] Object type restrictions on a first parameter of several Object.* functions are relaxed
https://bugs.webkit.org/show_bug.cgi?id=142937

Reviewed by Darin Adler.

Source/JavaScriptCore:

In ES6, Object type restrictions on a first parameter of several Object.* functions are relaxed.
In ES5 or prior, when a first parameter is not object type, these functions raise TypeError.
But now, several functions perform ToObject onto a non-object parameter.
And others behaves as if a parameter is a non-extensible ordinary object with no own properties.
It is described in ES6 Annex E.
Functions different from ES5 are following.

1. An attempt is make to coerce the argument using ToObject.
    Object.getOwnPropertyDescriptor
    Object.getOwnPropertyNames
    Object.getPrototypeOf
    Object.keys

2. Treated as if it was a non-extensible ordinary object with no own properties.
    Object.freeze
    Object.isExtensible
    Object.isFrozen
    Object.isSealed
    Object.preventExtensions
    Object.seal

* runtime/ObjectConstructor.cpp:
(JSC::ObjectConstructorGetPrototypeOfFunctor::operator()):
(JSC::objectConstructorGetPrototypeOf):
(JSC::objectConstructorGetOwnPropertyDescriptor):
(JSC::objectConstructorGetOwnPropertyNames):
(JSC::objectConstructorKeys):
(JSC::objectConstructorSeal):
(JSC::objectConstructorFreeze):
(JSC::objectConstructorPreventExtensions):
(JSC::objectConstructorIsSealed):
(JSC::objectConstructorIsFrozen):
(JSC::objectConstructorIsExtensible):
* tests/stress/object-freeze-accept-non-object.js: Added.
* tests/stress/object-get-own-property-descriptor-perform-to-object.js: Added.
(canary):
* tests/stress/object-get-own-property-names-perform-to-object.js: Added.
(compare):
* tests/stress/object-get-prototype-of-perform-to-object.js: Added.
* tests/stress/object-is-extensible-accept-non-object.js: Added.
* tests/stress/object-is-frozen-accept-non-object.js: Added.
* tests/stress/object-is-sealed-accept-non-object.js: Added.
* tests/stress/object-keys-perform-to-object.js: Added.
(compare):
* tests/stress/object-prevent-extensions-accept-non-object.js: Added.
* tests/stress/object-seal-accept-non-object.js: Added.

LayoutTests:

Several LayoutTests are updated with the new behavior.
Test cases in IETestCenter contain test for the outdated behavior.
But IETestCenter is brought from the external source, so instead of modifying it,
we mark them as Failure in TestExpectations.

* TestExpectations:
* js/dom/getOwnPropertyDescriptor-expected.txt:
* js/prototypes-expected.txt:
* js/resources/getOwnPropertyDescriptor.js:
* js/script-tests/prototypes.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182189 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[New Block-Inside-Inline Model] Make sure line breaks occur before and after the...
hyatt@apple.com [Tue, 31 Mar 2015 17:42:24 +0000 (17:42 +0000)]
[New Block-Inside-Inline Model] Make sure line breaks occur before and after the anonymous inline-block.
https://bugs.webkit.org/show_bug.cgi?id=143238.

Reviewed by Dean Jackson.

Source/WebCore:

Added fast/blocks/inside-inlines/breaking-behavior.html (and new-model equivalent).

* rendering/InlineFlowBox.h:
(WebCore::InlineFlowBox::InlineFlowBox):
(WebCore::InlineFlowBox::hasAnonymousInlineBlock):
(WebCore::InlineFlowBox::setHasAnonymousInlineBlock):
Add a bit to root line boxes (stored in the InlineFlowBox since it has free bits) that indicates whether
or not a line is an anonymous inline-block line.

* rendering/RenderBlockFlow.cpp:
(WebCore::RenderBlockFlow::computeInlinePreferredLogicalWidths):
Patch the preferred logical width computation to break both before and after an anonymous inline-block.
Also make sure to strip trailing spaces from the line before the anonymous inline-block.

* rendering/RenderBlockLineLayout.cpp:
(WebCore::RenderBlockFlow::constructLine):
Patch line construction so that no intermediate boxes are constructed between the InlineBox for the anonymous
inline-block and the root line box.

* rendering/RootInlineBox.cpp:
(WebCore::RootInlineBox::ascentAndDescentForBox):
The height of a "line" that holds an anonymous inline-block should ignore line-box-contain and also make sure
that the margins of the replaced element are "outside" the line, since those margins will collapse.

* rendering/line/BreakingContext.h:
(WebCore::BreakingContext::handleReplaced):
Patch handleReplaced to make sure breaks occur both before and after an anonymous inline-block. In the case of
a break after the block, we setPreviousLineBrokeCleanly to true so that <br>s that follow the anonymous inline-block
are respected.

LayoutTests:

* fast/block/inside-inlines/breaking-behavior-expected.html: Added.
* fast/block/inside-inlines/breaking-behavior.html: Added.
* fast/block/inside-inlines/new-model/breaking-behavior-expected.html: Added.
* fast/block/inside-inlines/new-model/breaking-behavior.html: Added.
* fast/block/inside-inlines/new-model/empty-block.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182188 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[EFL] Add nullptr check to GraphicsContext3D::makeContextCurrent()
ossy@webkit.org [Tue, 31 Mar 2015 17:32:46 +0000 (17:32 +0000)]
[EFL] Add nullptr check to GraphicsContext3D::makeContextCurrent()
https://bugs.webkit.org/show_bug.cgi?id=143214

Reviewed by Žan Doberšek.

* platform/graphics/efl/GraphicsContext3DEfl.cpp:
(WebCore::GraphicsContext3D::makeContextCurrent):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182187 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: add 2D/WebGL canvas instrumentation infrastructure
commit-queue@webkit.org [Tue, 31 Mar 2015 17:23:12 +0000 (17:23 +0000)]
Web Inspector: add 2D/WebGL canvas instrumentation infrastructure
https://bugs.webkit.org/show_bug.cgi?id=137278

Patch by Matt Baker <mattbaker@apple.com> on 2015-03-31
Reviewed by Timothy Hatcher.

Source/JavaScriptCore:

Added Canvas protocol which defines types used by InspectorCanvasAgent.

* CMakeLists.txt:
* DerivedSources.make:
* inspector/protocol/Canvas.json: Added.

* inspector/scripts/codegen/generator.py:
(Generator.stylized_name_for_enum_value):
Added special handling for 2D (always uppercase) and WebGL (rename mapping) enum strings.

Source/WebCore:

Added backend agent for canvas inspection. A canvas is instrumented once its CanvasRenderingContext has been
created. WebGLRenderingContext is instrumented to track the lifetime of shader program objects.

To instrument off-screen rendering contexts (CSS canvases and detached canvas elements), the canvas agent must
track all CanvasRenderingContexts as they are created, even in the absense of the frontend. As far as I know,
there is no practical way to identify rendering contexts belonging to a frame that are not in the DOM. In the
absence of the inspector frontend the agent does only the minimum required to track the lifetime of rendering
contexts and their resources, to avoid creating a negative performance impact.

Tests: inspector/canvas/canvas-add-remove-events.html
       inspector/canvas/canvas-context-attributes.html
       inspector/canvas/canvas-css-name.html

* CMakeLists.txt:
* WebCore.xcodeproj/project.pbxproj:

* dom/Document.cpp:
(WebCore::Document::getCSSCanvasElement):
Instrumentation for CSS canvases.

* html/HTMLCanvasElement.cpp:
(WebCore::HTMLCanvasElement::getContext):
Instrumentation for context creation.

* html/canvas/WebGLRenderingContextBase.cpp:
(WebCore::WebGLRenderingContextBase::attachShader):
(WebCore::WebGLRenderingContextBase::createProgram):
(WebCore::WebGLRenderingContextBase::deleteProgram):
(WebCore::WebGLRenderingContextBase::detachShader):
Instrumentation for programs and shaders.

* inspector/InspectorAllInOne.cpp:

* inspector/InspectorCanvasAgent.cpp: Added.
(WebCore::InspectorCanvasAgent::InspectorCanvasAgent):
(WebCore::InspectorCanvasAgent::didCreateFrontendAndBackend):
(WebCore::InspectorCanvasAgent::willDestroyFrontendAndBackend):
(WebCore::InspectorCanvasAgent::frameNavigated):
(WebCore::InspectorCanvasAgent::didCreateCSSCanvas):
(WebCore::InspectorCanvasAgent::didCreateCanvasRenderingContext):
(WebCore::InspectorCanvasAgent::didAttachShader):
(WebCore::InspectorCanvasAgent::didDetachShader):
(WebCore::InspectorCanvasAgent::didCreateProgram):
(WebCore::InspectorCanvasAgent::didDeleteProgram):
(WebCore::InspectorCanvasAgent::getCanvases):
(WebCore::InspectorCanvasAgent::canvasDestroyed):
(WebCore::InspectorCanvasAgent::canvasDestroyedTimerFired):
(WebCore::InspectorCanvasAgent::reset):
(WebCore::InspectorCanvasAgent::getCanvasEntry):
(WebCore::InspectorCanvasAgent::getProgramEntry):
(WebCore::InspectorCanvasAgent::removeShaderFromShaderMap):
(WebCore::InspectorCanvasAgent::contextTypeJson):
(WebCore::InspectorCanvasAgent::buildObjectForCanvas):
* inspector/InspectorCanvasAgent.h: Added.
New backend agent.

* inspector/InspectorController.cpp:
(WebCore::InspectorController::InspectorController):
Support for new backend agent.

* inspector/InspectorInstrumentation.cpp:
(WebCore::InspectorInstrumentation::didCommitLoadImpl):
(WebCore::InspectorInstrumentation::didCreateCSSCanvasImpl):
(WebCore::InspectorInstrumentation::didCreateCanvasRenderingContextImpl):
(WebCore::InspectorInstrumentation::didAttachShaderImpl):
(WebCore::InspectorInstrumentation::didDetachShaderImpl):
(WebCore::InspectorInstrumentation::didCreateProgramImpl):
(WebCore::InspectorInstrumentation::didDeleteProgramImpl):
(WebCore::InspectorInstrumentation::instrumentingAgentsForRenderingContext):
Plumbing for canvas instrumentation.

* inspector/InspectorInstrumentation.h:
(WebCore::InspectorInstrumentation::didCreateCSSCanvas):
(WebCore::InspectorInstrumentation::didCreateCanvasRenderingContext):
(WebCore::InspectorInstrumentation::didAttachShader):
(WebCore::InspectorInstrumentation::didDetachShader):
(WebCore::InspectorInstrumentation::didCreateProgram):
(WebCore::InspectorInstrumentation::didDeleteProgram):

* inspector/InstrumentingAgents.cpp:
(WebCore::InstrumentingAgents::InstrumentingAgents):
(WebCore::InstrumentingAgents::reset):
* inspector/InstrumentingAgents.h:
(WebCore::InstrumentingAgents::inspectorCanvasAgent):
(WebCore::InstrumentingAgents::setInspectorCanvasAgent):
Support for new backend agent.

Source/WebInspectorUI:

Added models, views, and controller classes for 2D and WebGL canvas inspection. Each canvas is shown in the
Resources navigation sidebar under its parent frame. Shader programs are displayed as child nodes of
their respective canvas. Canvases will get an associated content view and details sidebar in a later patch
(see https://bugs.webkit.org/show_bug.cgi?id=138941).

Shader programs will get an associated content view for editing shader source in a later patch (see
https://bugs.webkit.org/show_bug.cgi?id=138593). Individual shaders are not shown in the Resource navigation
sidebar, and at this time there are no plans to instrument shaders that aren't attached to a program.

* Localizations/en.lproj/localizedStrings.js:
* UserInterface/Base/Main.js:
(WebInspector.loaded):
(WebInspector.sidebarPanelForRepresentedObject):
* UserInterface/Base/Test.js:
(WebInspector.loaded):
* UserInterface/Main.html:
* UserInterface/Test.html:
Updated for new canvas/shader program types.

* UserInterface/Controllers/CanvasManager.js: Added.
(WebInspector.CanvasManager):
(WebInspector.CanvasManager.prototype.canvasesForFrame):
(WebInspector.CanvasManager.prototype.canvasAdded.set this):
(WebInspector.CanvasManager.prototype.canvasAdded):
(WebInspector.CanvasManager.prototype.canvasRemoved):
(WebInspector.CanvasManager.programDeleted.get console):
(WebInspector.CanvasManager.prototype.programCreated.get console):
(WebInspector.CanvasManager.prototype.programCreated):
(WebInspector.CanvasManager.prototype._mainResourceDidChange):
Frontend controller for canvases and their shader programs.

* UserInterface/Images/Canvas.svg: Added.
* UserInterface/Images/DocumentGL.png: Added.
* UserInterface/Images/DocumentGL@2x.png: Added.
New art for canvas and shader program tree elements.

* UserInterface/Models/Canvas.js: Added.
(WebInspector.Canvas):
(WebInspector.Canvas.prototype.set fromPayload):
(WebInspector.Canvas.displayNameForContextType):
(WebInspector.Canvas.resetUniqueDisplayNameNumbers):
(WebInspector.Canvas.prototype.get id):
(WebInspector.Canvas.prototype.get parentFrame):
(WebInspector.Canvas.prototype.get name):
(WebInspector.Canvas.prototype.get cssCanvas):
(WebInspector.Canvas.prototype.get contextType):
(WebInspector.Canvas.prototype.get contextAttributes):
(WebInspector.Canvas.prototype.get programs):
(WebInspector.Canvas.prototype.get displayName):
(WebInspector.Canvas.programForId):
(WebInspector.Canvas.get programWasCreated.set this):
(WebInspector.Canvas.prototype.get programWasCreated):
(WebInspector.Canvas.prototype.programWasDeleted):
(WebInspector.Canvas.prototype.saveIdentityToCookie):
Model for DOM or CSS canvas (2D or WebGL).

* UserInterface/Models/ShaderProgram.js: Added.
(WebInspector.ShaderProgram):
(WebInspector.ShaderProgram.prototype.get id):
(WebInspector.ShaderProgram.prototype.get canvas):
(WebInspector.ShaderProgram.prototype.get displayName):
(WebInspector.ShaderProgram.prototype.saveIdentityToCookie):
(WebInspector.ShaderProgram.prototype.updateCanvas):
* UserInterface/Models/WebGLContextAttributes.js: Added.
(WebInspector.WebGLContextAttributes):
(WebInspector.WebGLContextAttributes.fromPayload):
(WebInspector.WebGLContextAttributes.prototype.get alpha):
(WebInspector.WebGLContextAttributes.prototype.get depth):
(WebInspector.WebGLContextAttributes.prototype.get stencil):
(WebInspector.WebGLContextAttributes.prototype.get antialias):
(WebInspector.WebGLContextAttributes.prototype.get premultipliedAlpha):
(WebInspector.WebGLContextAttributes.prototype.get preserveDrawingBuffer):
Model for WebGL canvas context attributes.

* UserInterface/Protocol/CanvasObserver.js: Added.
(WebInspector.CanvasObserver.prototype.canvasAdded):
(WebInspector.CanvasObserver.prototype.canvasRemoved):
(WebInspector.CanvasObserver.prototype.programCreated):
(WebInspector.CanvasObserver.prototype.programDeleted):
(WebInspector.CanvasObserver):
Model for WebGL canvas shader program.

* UserInterface/Views/CanvasTreeElement.js: Added.
(WebInspector.CanvasTreeElement.validateRepresentedObject):
(WebInspector.CanvasTreeElement.countChildren):
(WebInspector.CanvasTreeElement):
(WebInspector.CanvasTreeElement.prototype.onexpand):
(WebInspector.CanvasTreeElement.prototype.oncollapse):
(WebInspector.CanvasTreeElement.prototype.onpopulate):
(WebInspector.CanvasTreeElement.prototype._programWasCreated):
(WebInspector.CanvasTreeElement.prototype._programWasDeleted):
Folderized tree element for canvases and their child objects (shader programs).

* UserInterface/Views/FrameTreeElement.js:
(WebInspector.FrameTreeElement.prototype.onattach):
(WebInspector.FrameTreeElement.prototype.ondetach):
(WebInspector.FrameTreeElement.prototype.onpopulate):
(WebInspector.FrameTreeElement.prototype._canvasesAvailable):
(WebInspector.FrameTreeElement.prototype._canvasWasAdded):
(WebInspector.FrameTreeElement.prototype._canvasWasRemoved):
Updated to support canvas tree elements.

* UserInterface/Views/ResourceIcons.css:
(.canvas-icon .icon):
(.shader-program-icon .icon):
Styles for new canvas and shader program icons.

* UserInterface/Views/ResourceSidebarPanel.js:
(WebInspector.ResourceSidebarPanel.prototype._treeElementSelected):
Updated for new tree element types.

* UserInterface/Views/ShaderProgramTreeElement.js: Added.
(WebInspector.ShaderProgramTreeElement):
Tree element for shader programs. Shown as children of CanvasTreeElement.

* WebInspectorUI.vcxproj/WebInspectorUI.vcxproj:
* WebInspectorUI.vcxproj/WebInspectorUI.vcxproj.filters:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182186 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoPerf dashboard should show relative change in values
rniwa@webkit.org [Tue, 31 Mar 2015 17:17:07 +0000 (17:17 +0000)]
Perf dashboard should show relative change in values
https://bugs.webkit.org/show_bug.cgi?id=143252

Reviewed by Antti Koivisto.

When a range of values are selected, show the percentage difference between the start and the end
in addition to the absolute value difference. When a single point is selected, show the relative
difference with respect to the previous point. Use two significant figures and always show plus sign
when the difference is positive.

* public/v2/app.js: Compute and format the relative difference.
* public/v2/chart-pane.css: Don't let commits view shrink itself when they're all collapsed.
* public/v2/index.html: Show the relative difference.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182185 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoREGRESSION(r180000): Changing moving average or enveloping strategy doesn't update...
rniwa@webkit.org [Tue, 31 Mar 2015 17:16:07 +0000 (17:16 +0000)]
REGRESSION(r180000): Changing moving average or enveloping strategy doesn't update the graph
https://bugs.webkit.org/show_bug.cgi?id=143254

Reviewed by Antti Koivisto.

The bug was caused by App.Pane no longer replacing 'chartData' property when updating the moving average
or the enveloping values. Fixed the bug by creating a new chartData object when the strategy is changed
so that the interactive chart component will observe a change to 'chartData'.

* public/v2/app.js:
(App.Pane._movingAverageOrEnvelopeStrategyDidChange): Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182184 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoGTK+ Gardening 30th March
commit-queue@webkit.org [Tue, 31 Mar 2015 12:35:25 +0000 (12:35 +0000)]
GTK+ Gardening 30th March
https://bugs.webkit.org/show_bug.cgi?id=143220

Unreviewed.

Patch by Marcos Chavarría Teijeiro <chavarria1991@gmail.com> on 2015-03-31

* platform/gtk/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182183 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, revert accidental change by r182138.
ossy@webkit.org [Tue, 31 Mar 2015 12:23:04 +0000 (12:23 +0000)]
Unreviewed, revert accidental change by r182138.

* Scripts/generate-forwarding-headers.pl:
(createForwardingHeadersForFramework):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182182 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agohttp/tests/cache/disk-cache/disk-cache-validation-back-navigation-policy.html failing...
antti@apple.com [Tue, 31 Mar 2015 10:59:24 +0000 (10:59 +0000)]
http/tests/cache/disk-cache/disk-cache-validation-back-navigation-policy.html failing on Mavericks wk2 debug
https://bugs.webkit.org/show_bug.cgi?id=143258

The test relies on resource priorities. Those don't roundtrip when serializing via NSURLRequest on Mavericks due
to ResourceRequest::resourcePrioritiesEnabled() returning false.

It is bit surprising it works on release but not on debug. I suspect some debug code forces NSURLRequest construction
and so we end up on the roundtrip code path.

* platform/mac-mavericks/TestExpectations: Added.

    Skip the test on Mavericks.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182181 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Streams API] Implement a barebone ReadableStreamReader interface
youenn.fablet@crf.canon.fr [Tue, 31 Mar 2015 10:09:56 +0000 (10:09 +0000)]
[Streams API] Implement a barebone ReadableStreamReader interface
https://bugs.webkit.org/show_bug.cgi?id=142866

Reviewed by Benjamin Poulain.

Source/WebCore:

This patch implements a barebone ReadableStreamReader as described in https://streams.spec.whatwg.org/#reader-class.
It implements the IDL without implementing most of the functionality behind.
It implements ReadableStreamReader JS constructor and the link with ReadableStream.getReader:
- Raise an error if stream is already locked
- Set the stream reader to the newly created ReadableStreamReader instance.

Test: streams/readablestreamreader-constructor.html

* CMakeLists.txt: Adding ReadableStreamReader files.
* DerivedSources.cpp: Ditto.
* DerivedSources.make: Ditto.
* Modules/streams/ReadableStream.cpp:
(WebCore::ReadableStream::createReader): Used by JS binding to get access to the reader.
* Modules/streams/ReadableStream.h: Declaration of createReader, reader and lock.
(WebCore::ReadableStream::reader):
(WebCore::ReadableStream::lock):
* Modules/streams/ReadableStream.idl: Cleaning-up the IDL.
* Modules/streams/ReadableStreamReader.cpp: Added.
(WebCore::ReadableStreamReader::create):
(WebCore::ReadableStreamReader::ReadableStreamReader):
(WebCore::ReadableStreamReader::~ReadableStreamReader):
(WebCore::ReadableStreamReader::closed):
(WebCore::ReadableStreamReader::activeDOMObjectName):
(WebCore::ReadableStreamReader::canSuspend):
* Modules/streams/ReadableStreamReader.h: Added.
(WebCore::ReadableStreamReader::stream):
* Modules/streams/ReadableStreamReader.idl: Added.
* WebCore.vcxproj/WebCore.vcxproj:
* WebCore.vcxproj/WebCore.vcxproj.filters:
* WebCore.xcodeproj/project.pbxproj:
* bindings/js/JSBindingsAllInOne.cpp:
* bindings/js/JSReadableStreamCustom.cpp:
(WebCore::JSReadableStream::getReader):
* bindings/js/JSReadableStreamReaderCustom.cpp: Added.
(WebCore::JSReadableStreamReader::read):
(WebCore::getOrCreatePromiseDeferredFromObject):
(WebCore::closedPromiseSlotName):
(WebCore::JSReadableStreamReader::closed):
(WebCore::JSReadableStreamReader::cancel):
(WebCore::JSReadableStreamReader::releaseLock):
(WebCore::constructJSReadableStreamReader):

LayoutTests:

Adding tests to check API, constructor and link with ReadableStream.getReader.
Rebasing of global-constructors-attributes.html to include ReadableStreamReader.

* js/dom/global-constructors-attributes-expected.txt:
* platform/efl/js/dom/global-constructors-attributes-expected.txt:
* platform/gtk/js/dom/global-constructors-attributes-expected.txt:
* platform/mac-mavericks/js/dom/global-constructors-attributes-expected.txt:
* platform/mac/js/dom/global-constructors-attributes-expected.txt:
* platform/win/js/dom/global-constructors-attributes-expected.txt:
* streams/readablestreamreader-constructor-expected.txt: Added.
* streams/readablestreamreader-constructor.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182180 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agounreviewed config fix after r182178.
lforschler@apple.com [Tue, 31 Mar 2015 10:05:29 +0000 (10:05 +0000)]
unreviewed config fix after r182178.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182179 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[EFL] Add AArch64 builder and tester bot to build.webkit.org
ossy@webkit.org [Tue, 31 Mar 2015 09:19:59 +0000 (09:19 +0000)]
[EFL] Add AArch64 builder and tester bot to build.webkit.org
https://bugs.webkit.org/show_bug.cgi?id=143092

Reviewed by Gyuyoung Kim.

* BuildSlaveSupport/build.webkit.org-config/config.json:
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
(Factory.__init__):
* BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182178 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[buildbot] clean-build script should remove untracked files and revert local changes too
youenn.fablet@crf.canon.fr [Tue, 31 Mar 2015 08:58:31 +0000 (08:58 +0000)]
[buildbot] clean-build script should remove untracked files and revert local changes too
https://bugs.webkit.org/show_bug.cgi?id=142400

Reviewed by Ryosuke Niwa.

This patch cleans the WebKit folder by reverting tracked files changes and deleting SCM untracked files, including SCM ignored files.
A helper routine SCM.discard_untracked_files is added for that purpose.

* BuildSlaveSupport/clean-build:
(main): Making call to Scripts/clean-webkit
* Scripts/clean-webkit: Added.
(main): Revert changes and delete untracked files.
* Scripts/webkitpy/common/checkout/scm/scm.py:
(SCM.discard_untracked_files): Helper function to discard untracked files or folders found by SCM.
* Scripts/webkitpy/common/checkout/scm/scm_unittest.py:
(SCMTest._shared_test_discard_untracked_files):
(test_discard_untracked_files): Tests that untracked file and untracked folder get discarded correctly.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182177 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed. Fix GTK+ build with REDIRECTED_XCOMPOSITE_WINDOW disabled in X11 platform.
carlosgc@webkit.org [Tue, 31 Mar 2015 08:37:04 +0000 (08:37 +0000)]
Unreviewed. Fix GTK+ build with REDIRECTED_XCOMPOSITE_WINDOW disabled in X11 platform.

Also fix some unused parameter warnings when
REDIRECTED_XCOMPOSITE_WINDOW is disabled.

* UIProcess/API/gtk/WebKitWebViewBase.cpp:
(webkitWebViewRenderAcceleratedCompositingResults):
(resizeWebKitWebViewBaseFromAllocation):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182176 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[GTK] DnD icon/widget has odd background
carlosgc@webkit.org [Tue, 31 Mar 2015 07:49:54 +0000 (07:49 +0000)]
[GTK] DnD icon/widget has odd background
https://bugs.webkit.org/show_bug.cgi?id=143217

Reviewed by Martin Robinson.

Source/WebCore:

Remove DragIcon class since it's no longer needed with GTK+3 and
the GTK+2 code there is unused. GTK+ knows what to do with a cairo
surface, I guess we migrated the GTK+2 code to GTK+3 without
realizing that using the surface was enough.

* PlatformGTK.cmake:
* platform/gtk/DragIcon.cpp: Removed.
* platform/gtk/DragIcon.h: Removed.

Source/WebKit2:

Use gtk_drag_set_icon_surface() to set the drag icon image,
instead of DragIcon class.

* UIProcess/API/gtk/WebKitWebView.cpp:
* UIProcess/gtk/DragAndDropHandler.cpp:
(WebKit::DragAndDropHandler::startDrag):
* UIProcess/gtk/DragAndDropHandler.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182175 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoPython SCM should be able to retrieve untracked files
youenn.fablet@crf.canon.fr [Tue, 31 Mar 2015 07:34:11 +0000 (07:34 +0000)]
Python SCM should be able to retrieve untracked files
https://bugs.webkit.org/show_bug.cgi?id=143135

Reviewed by Ryosuke Niwa.

This patch introduces SCM.untracked_files to retrieve a list of untracked files.
Ignored files may be included or not in this list according boolean passed to the function.

* Scripts/webkitpy/common/checkout/scm/git.py:
(Git.untracked_files):
* Scripts/webkitpy/common/checkout/scm/scm.py:
(SCM.untracked_files):
* Scripts/webkitpy/common/checkout/scm/scm_unittest.py:
(SCMTest._shared_test_untracked_files):
(test_untracked_files):
(GitSVNTest.test_untracked_files):
* Scripts/webkitpy/common/checkout/scm/svn.py:
(SVN.untracked_files):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182174 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r182172.
commit-queue@webkit.org [Tue, 31 Mar 2015 03:27:01 +0000 (03:27 +0000)]
Unreviewed, rolling out r182172.
https://bugs.webkit.org/show_bug.cgi?id=143250

Since this class is an API object, we don't know for sure when
it will be destroyed. (Requested by kling on #webkit).

Reverted changeset:

"WebBackForwardList doesn't need to manually clear the
snapshot of an entry when removing it."
https://bugs.webkit.org/show_bug.cgi?id=143231
http://trac.webkit.org/changeset/182172

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182173 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWebBackForwardList doesn't need to manually clear the snapshot of an entry when remov...
akling@apple.com [Tue, 31 Mar 2015 03:02:41 +0000 (03:02 +0000)]
WebBackForwardList doesn't need to manually clear the snapshot of an entry when removing it.
<https://webkit.org/b/143231>

Reviewed by Tim Horton.

Now that we don't keep WebBackForwardListItem objects alive forever, it's not necessary
to clear the snapshot when removing them from the WebBackForwardList, since the entire
object will be deleted shortly after that anyway.

* UIProcess/WebBackForwardList.cpp:
(WebKit::WebBackForwardList::didRemoveItem):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182172 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoExtending null should set __proto__ to null
rniwa@webkit.org [Tue, 31 Mar 2015 01:59:31 +0000 (01:59 +0000)]
Extending null should set __proto__ to null
https://bugs.webkit.org/show_bug.cgi?id=142882

Reviewed by Geoffrey Garen and Benjamin Poulain.

Source/JavaScriptCore:

Set Derived.prototype.__proto__ to null when extending null.

* bytecompiler/NodesCodegen.cpp:
(JSC::ClassExprNode::emitBytecode):

LayoutTests:

Added more test cases for extending from null, and added checks for prototypes.

Also rebaselined existing tests.

* js/class-syntax-call-expected.txt:
* js/class-syntax-extends-expected.txt:
* js/class-syntax-super-expected.txt:
* js/script-tests/class-syntax-call.js:
* js/script-tests/class-syntax-extends.js:
* js/script-tests/class-syntax-super.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182171 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix failing v8-deltablue.js for ARM
msaboff@apple.com [Tue, 31 Mar 2015 01:21:21 +0000 (01:21 +0000)]
Fix failing v8-deltablue.js for ARM
https://bugs.webkit.org/show_bug.cgi?id=138392

Rubber-stamped by Mark Lam.

Re-enable this test for ARM32 on iOS as the tests now run without issues.

* tests/v8-v6/v8-deltablue.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182170 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed build fix after r182168.
achristensen@apple.com [Tue, 31 Mar 2015 01:09:08 +0000 (01:09 +0000)]
Unreviewed build fix after r182168.

* contentextensions/ContentExtensionsBackend.cpp:
(WebCore::ContentExtensions::ContentExtensionsBackend::actionsForResourceLoad):
Explicitly cast to unsigned.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182169 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Content Extensions] Correctly handle regular expressions matching everything
achristensen@apple.com [Tue, 31 Mar 2015 00:48:30 +0000 (00:48 +0000)]
[Content Extensions] Correctly handle regular expressions matching everything
https://bugs.webkit.org/show_bug.cgi?id=143235

Reviewed by Benjamin Poulain.

Source/WebCore:

Test: http/tests/contentextensions/whitelist.html

* contentextensions/CompiledContentExtension.cpp:
(WebCore::ContentExtensions::CompiledContentExtension::globalDisplayNoneSelectors):
Global actions can have non-css actions.  Only put the selectors into the list of selectors.
* contentextensions/ContentExtensionCompiler.cpp:
(WebCore::ContentExtensions::serializeSelector):
(WebCore::ContentExtensions::serializeActions):
Merge sequential css selectors with identical triggers (usually .*) into one action to reduce the number of actions.
(WebCore::ContentExtensions::compileRuleList):
Fail if a regular expression matches everything after ignore-previous-rules.
* contentextensions/ContentExtensionError.cpp:
(WebCore::ContentExtensions::contentExtensionErrorCategory):
* contentextensions/ContentExtensionError.h:
Add more failure cases.
* contentextensions/ContentExtensionRule.h:
(WebCore::ContentExtensions::Trigger::operator==):
Allow comparing of Triggers to determine if sequential triggers are equal.
* contentextensions/ContentExtensionsBackend.cpp:
(WebCore::ContentExtensions::ContentExtensionsBackend::actionsForResourceLoad):
Put non-css actions that match everything into the list of actions if ignore-previous-rules was not hit.
These actions will be out of order, but order only matters when determining if ignore-previous-rules, and this case is handled correctly.
* contentextensions/DFABytecodeInterpreter.cpp:
(WebCore::ContentExtensions::DFABytecodeInterpreter::actionsFromDFARoot):
(WebCore::ContentExtensions::DFABytecodeInterpreter::interpret):
Added an assertion that all actions that match everything should be in the first DFA root.
We should catch them all with URLFilterParser::MatchesEverything.

Tools:

* TestWebKitAPI/Tests/WebCore/ContentExtensions.cpp:
(TestWebKitAPI::checkCompilerError):
(TestWebKitAPI::TEST_F):
Test ContentExtensionErrors.

LayoutTests:

* http/tests/contentextensions/css-display-none.html:
* http/tests/contentextensions/css-display-none.html.json:
Test multiple selectors with triggers that match everything.
* http/tests/contentextensions/whitelist-expected.txt: Added.
* http/tests/contentextensions/whitelist.html: Added.
* http/tests/contentextensions/whitelist.html.json: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182168 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoREGRESSION (r181993): inspector-protocol/debugger/setBreakpoint-dfg-and-modify-local...
mark.lam@apple.com [Tue, 31 Mar 2015 00:43:49 +0000 (00:43 +0000)]
REGRESSION (r181993): inspector-protocol/debugger/setBreakpoint-dfg-and-modify-local.html crashes.
<https://webkit.org/b/143105>

Reviewed by Filip Pizlo.

Source/JavaScriptCore:

With r181993, the DFG and FTL may elide the storing of the scope register.  As a result,
on OSR exits from DFG / FTL frames where this elision has take place, we may get baseline
JIT frames that may have its scope register not set.  The Debugger's current implementation
which relies on the scope register is not happy about this.  For example, this results in a
crash in the layout test inspector-protocol/debugger/setBreakpoint-dfg-and-modify-local.html.

The fix is to disable inlining when the debugger is in use.  Also, we add Flush nodes to
ensure that the scope register value is flushed to the register in the stack frame.

* dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::ByteCodeParser):
(JSC::DFG::ByteCodeParser::setLocal):
(JSC::DFG::ByteCodeParser::flush):
- Add code to flush the scope register.
(JSC::DFG::ByteCodeParser::inliningCost):
- Pretend that all codeBlocks are too expensive to inline if the debugger is in use, thereby
  disabling inlining whenever the debugger is in use.
* dfg/DFGGraph.cpp:
(JSC::DFG::Graph::Graph):
* dfg/DFGGraph.h:
(JSC::DFG::Graph::hasDebuggerEnabled):
* dfg/DFGStackLayoutPhase.cpp:
(JSC::DFG::StackLayoutPhase::run):
- Update the DFG codeBlock's scopeRegister since it can be moved during stack layout.
* ftl/FTLCompile.cpp:
(JSC::FTL::mmAllocateDataSection):
- Update the FTL codeBlock's scopeRegister since it can be moved during stack layout.

LayoutTests:

* TestExpectations:
- Undid test skipped in r182072.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182167 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoDumpRenderTree should set NSWindowDisplayWithRunLoopObserver
ap@apple.com [Tue, 31 Mar 2015 00:33:19 +0000 (00:33 +0000)]
DumpRenderTree should set NSWindowDisplayWithRunLoopObserver
https://bugs.webkit.org/show_bug.cgi?id=143241
rdar://problem/20351297

Reviewed by Mark Lam.

* DumpRenderTree/mac/DumpRenderTree.mm: (setDefaultsToConsistentValuesForTesting):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182166 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[CMake][EFL] Remove redundant library link and include path
gyuyoung.kim@samsung.com [Tue, 31 Mar 2015 00:29:17 +0000 (00:29 +0000)]
[CMake][EFL] Remove redundant library link and include path
https://bugs.webkit.org/show_bug.cgi?id=143221

Reviewed by Csaba Osztrogonác.

Source/WTF:

* wtf/PlatformEfl.cmake: Remove unnecessary library link and include path.

Tools:

* ImageDiff/PlatformEfl.cmake: Remove unnecessary library link and include path.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182165 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAllow building on Windows without Cygwin
mmaxfield@apple.com [Mon, 30 Mar 2015 23:59:02 +0000 (23:59 +0000)]
Allow building on Windows without Cygwin
https://bugs.webkit.org/show_bug.cgi?id=143219

Reviewed by Brent Fulgham.

* AVFoundationSupport.py: Renamed from Source/WebCore/WebCore.vcxproj/MigrateScripts.
Move logic from DerivedSources.make to its own script
(lookFor):
* DerivedSources.make: Paths like /usr/bin don't exist on Windows.
Windows uses del instead of rm -f
Windows doesn't have printf
* WebCore.vcxproj/WebCoreGenerated.make: Pass the CC executable to
build-generated-files.pl
* WebCore.vcxproj/build-generated-files.pl: Export the CC executable as an environment
variable. Don't use /usr/bin.
* WebCore.vcxproj/migrate-scripts.pl: Deleted the MigrateScripts makefile. We don't
need a whole Makefile just to copy 5 files. Moved this logic into migrate-scripts.pl.
(copyFile):
* bindings/scripts/preprocessor.pm:
(applyPreprocessor): Support passing preprocessing flags to the Visual Studio
compiler.
* dom/make_names.pl: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182164 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoBuild fix.
ap@apple.com [Mon, 30 Mar 2015 23:54:01 +0000 (23:54 +0000)]
Build fix.

* platform/spi/mac/AVFoundationSPI.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182163 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[iOS] WebContent crashing at WebCore: WebCore::Range::collectSelectionRects.
enrica@apple.com [Mon, 30 Mar 2015 23:25:44 +0000 (23:25 +0000)]
[iOS] WebContent crashing at WebCore: WebCore::Range::collectSelectionRects.
https://bugs.webkit.org/show_bug.cgi?id=143234
<rdar://problem/18571345>

Reviewed by Tim Horton.

This is a speculative fix that adds a null check before referencing the range.
In both places where the check has been added the range returned by the call
that should create it could be null.

* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::platformEditorState):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182162 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Content Extensions] Flesh out the UserContentExtensionStore
commit-queue@webkit.org [Mon, 30 Mar 2015 23:22:09 +0000 (23:22 +0000)]
[Content Extensions] Flesh out the UserContentExtensionStore
https://bugs.webkit.org/show_bug.cgi?id=143123

Patch by Sam Weinig <sam@webkit.org> on 2015-03-30
Reviewed by Benjamin Poulain.

Source/WebCore:

- Adds a variant of getFileSize() that takes a PlatformFileHandle.
- Adds a new file system function, rename() which operates as the POSIX
  rename function does.

* platform/FileSystem.h:
Add new functions.

* platform/posix/FileSystemPOSIX.cpp:
(WebCore::renameFile):
(WebCore::getFileSize):
Add implementations.

* platform/gtk/FileSystemGtk.cpp:
(WebCore::renameFile):
(WebCore::getFileSize):
* platform/win/FileSystemWin.cpp:
(WebCore::getFileSize):
(WebCore::renameFile):
Add stubs.

Source/WebKit2:

* Shared/WebCompiledContentExtension.cpp:
(WebKit::WebCompiledContentExtension::createFromCompiledContentExtensionData):
* Shared/WebCompiledContentExtensionData.cpp:
(WebKit::WebCompiledContentExtensionData::encode):
(WebKit::WebCompiledContentExtensionData::decode):
* Shared/WebCompiledContentExtensionData.h:
(WebKit::WebCompiledContentExtensionData::WebCompiledContentExtensionData):
Switch the order in memory of actions and bytecode (and switch the order of the
variables as well, to keep it clear). This will become necessary when streaming
the data to disk, as actions are created before the bytecode and we would have to keep them
in memory until the bytecode was finished compiling if they didn't come before in the file.

* UIProcess/API/APIUserContentExtensionStore.h:
* UIProcess/API/APIUserContentExtensionStore.cpp:
(API::UserContentExtensionStore::defaultStore):
Add accessor for the processes default shared store.

(API::constructedPath):
Helper for constructing the path to a file in the store based on identifier.

(API::encodeContentExtensionMetaData):
(API::decodeContentExtensionMetaData):
Helpers for encoding/decoding the file metadata (version, bytecode size, actions size).

(API::openAndMapContentExtension):
Helper to open and map a filed back content extension.

(API::writeDataToFile):
Helper to write a Data object to a file.

(API::compiledToFile):
Helper to run the content extension compiler and write it to disk. It first
writes it to a temporary file and then does an atomic rename operation to put
the file in the final location. Doing this means that if the process crashes while
the compile is taking place, the partially written file won't end up in the cache,
but rather, will be cleaned up by the OS.

(API::UserContentExtensionStore::lookupContentExtension):
API entry point to lookup an extension that has been compiled to disk. On a background queue,
it attempts to open and map the extension (based on the identifier passed in) and return it
to the caller via the callback function passed in.

(API::UserContentExtensionStore::compileContentExtension):
API entry point to compile an extension and write it to store. On a background queue, it
compiles the file to disk and if successful, returns the memory mapped data via the callback
function passed in.

(API::UserContentExtensionStore::removeContentExtension):
API entry point to remove an extension from the store. On a background queue, it
unlinks the extension (based on the identifier passed in) and tells the caller it completed
the action vial the callback function passed in.

(API::UserContentExtensionStore::synchronousRemoveAllContentExtensions):
Helper for testing which synchronously unlinks all the files in the store.

(API::userContentExtensionStoreErrorCategory):
std::error_code adaptor for some generic errors that can happen in the store. We will want
to flesh these out further.

* UIProcess/API/Cocoa/APIUserContentExtensionStoreCocoa.mm: Added.
(API::UserContentExtensionStore::defaultStorePath):
Helper to get the platform specific path for the store.

* UIProcess/API/C/WKUserContentExtensionStoreRef.cpp:
(WKUserContentExtensionStoreGetTypeID):
Add ENABLE(CONTENT_EXTENSIONS) guards.

* UIProcess/API/Cocoa/_WKUserContentExtensionStore.h:
* UIProcess/API/Cocoa/_WKUserContentExtensionStore.mm:
(-[_WKUserContentExtensionStore dealloc]):
(+[_WKUserContentExtensionStore defaultStore]):
(-[_WKUserContentExtensionStore compileContentExtensionForIdentifier:encodedContentExtension:completionHandler:]):
(-[_WKUserContentExtensionStore lookupContentExtensionForIdentifier:completionHandler:]):
(-[_WKUserContentExtensionStore removeContentExtensionForIdentifier:completionHandler:]):
(-[_WKUserContentExtensionStore _removeAllContentExtensions]):
(-[_WKUserContentExtensionStore init]): Deleted.
* UIProcess/API/Cocoa/_WKUserContentExtensionStoreInternal.h:
* UIProcess/API/Cocoa/_WKUserContentExtensionStorePrivate.h: Added.
SPI wrappers for the store.

* WebKit2.xcodeproj/project.pbxproj:
Add new files.

Tools:

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKit2Cocoa/_WKUserContentExtensionStore.mm: Added.
Add tests for _WKUserContentExtensionStore.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@182161 268f45cc-cd09-0410-ab3c-d52691b4dbfc