Context Navigation

It is redundant with RTCController::m_shouldFilterICECandidates.
Manually tested on webrtc test samples.
This code path is currently untested by rwt as we hook it to getUserMedia which is mocked in rwt.
Additional testing will be added as a follow-up patch.

WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::disableICECandidateFiltering):
(WebKit::WebPage::enableICECandidateFiltering):

WebProcess/WebPage/WebPage.h:

9:08 PM Changeset in webkit [214459] by jfbastien@apple.com

2 edits in trunk/Source/WebKit

make build fix for WebDefaultUIDelegate.m
https://bugs.webkit.org/show_bug.cgi?id=170163

Unreviewed build fix.

PlatformMac.cmake: WebDefaultUIDelegate.m was renamed

8:55 PM Changeset in webkit [214458] by Lucas Forschler

1 edit in trunk/Tools/ChangeLog

Another test v5.

8:50 PM Changeset in webkit [214457] by Lucas Forschler

1 edit in trunk/Tools/ChangeLog

Another test v4.

8:41 PM Changeset in webkit [214456] by Lucas Forschler

1 edit in trunk/Tools/ChangeLog

Another test v3.

8:06 PM Changeset in webkit [214455] by Wenson Hsieh

2 edits in trunk/Source/WebKit2

[WK2] UIKit assertion when rotating device after dismissing popover
https://bugs.webkit.org/show_bug.cgi?id=170159
<rdar://problem/31283829>

Reviewed by Tim Horton.

Ensure that when we are finished presenting a WKActionSheet, we unregister with the shared notifications center.

UIProcess/ios/WKActionSheet.mm:

(-[WKActionSheet dealloc]):
(-[WKActionSheet _cleanup]):
(-[WKActionSheet doneWithSheet:]):

8:05 PM Changeset in webkit [214454] by Lucas Forschler

1 edit in trunk/Tools/ChangeLog

Another test v2.

8:03 PM Changeset in webkit [214453] by Lucas Forschler

1 edit in trunk/Tools/ChangeLog

Another test.

7:53 PM Changeset in webkit [214452] by Lucas Forschler

1 edit in trunk/Tools/ChangeLog

<rdar://problem/30949128>
Update build.webkit.org to use port 16000 for PBChangeSource.

Reviewed by Alexey Proskuryakov.

BuildSlaveSupport/build.webkit.org-config/master.cfg:

7:43 PM Changeset in webkit [214451] by Lucas Forschler

1 edit in trunk/Tools/ChangeLog

Test commit to verify post-commit hook.

7:23 PM Changeset in webkit [214450] by commit-queue@webkit.org

42 edits
8 adds in trunk

REGRESSION(213764): Large images should not be decoded asynchronously when they are drawn on a canvas
https://bugs.webkit.org/show_bug.cgi?id=169771

Patch by Said Abou-Hallawa <sabouhallawa@apple.com> on 2017-03-27
Reviewed by Simon Fraser.

Source/WebCore:

Sometimes we have to draw the image immediately like when a canvas calls
drawImage. In this case we have to decode the image synchronously to guarantee
the drawing. Other times we need to decode with the native size of the image.
The animated images have to be decoded with native size always. Otherwise
the frame cache will be messed up if the same image is animated with different
sizes. Currently we always decode asynchronously with sizeForDrawing. We need
to decouple the decoding mode from the sizeForDrawing.

This patch introduce the DecodingOptions class which can store and compare the
following four cases:

-- Synchronous: The frame has be decoded with native size only.
-- Asynchronous + anySize: This is passed from the Image::draw() callers.
-- Asynchronous + fullSize: The image has to be decoded with its full size.
-- Asynchronous + sizeForDrawing: The image can be decoded with sizeForDrawing unless
it was decoded with either a full size or sizeForDrawing which is larger than the
requested sizeForDrawing.

A new argument of type DecodingMode will be added to Image::draw() function.
Only when the drawing comes from the render tree, it will be Asynchronous.
Otherwise it will be Synchronous.

Tests: fast/images/animated-image-different-dest-size.html

fast/images/async-image-background-image.html
fast/images/async-image-canvas-draw-image.html

WebCore.xcodeproj/project.pbxproj:
platform/graphics/BitmapImage.cpp:

(WebCore::BitmapImage::frameImageAtIndexCacheIfNeeded): Gets the frame image, cache it synchronously if
the current one is invalid. frameImageAtIndex() returns whatever stored in the cache.
(WebCore::BitmapImage::nativeImage): Call frameImageAtIndexCacheIfNeeded() instead of frameImageAtIndex().
(WebCore::BitmapImage::nativeImageForCurrentFrame): Ditto.
(WebCore::BitmapImage::nativeImageOfSize): Ditto.
(WebCore::BitmapImage::framesNativeImages): Ditto.
(WebCore::BitmapImage::draw): Change the logic to do the following:
-- The animated image has to be decoded with its full size.
-- The animated image expects the current frame to be ready for drawing.
-- The large image decoding does not need to call internalStartAnimation().
-- The large image has to request async image decoding but draw the current one if it exists.
(WebCore::BitmapImage::drawPattern): Draw the pattern synchronously.
(WebCore::BitmapImage::shouldUseAsyncDecodingForLargeImages): Delete the call to shouldUseAsyncDecodingForTesting()
since it is only applied for animated images.
(WebCore::BitmapImage::shouldUseAsyncDecodingForAnimatedImages): Call shouldUseAsyncDecodingForAnimatedImageForTesting().
(WebCore::BitmapImage::internalStartAnimation): Request decoding with the full size.
(WebCore::BitmapImage::advanceAnimation): Call shouldUseAsyncDecodingForAnimatedImageForTesting().
(WebCore::BitmapImage::internalAdvanceAnimation): Assert the current frame is not being decoding asynchronously for any size.
(WebCore::BitmapImage::frameImageAtIndex): Deleted. Moved to the header file but with a new purpose: return
the current frame from the frame cache as is; do not cache a new one.
(WebCore::BitmapImage::shouldUseAsyncDecodingForLargeImage): Deleted. Function was renamed to shouldUseAsyncDecodingForLargeImages.
(WebCore::BitmapImage::shouldUseAsyncDecodingForAnimatedImage): Deleted. Function was renamed to shouldUseAsyncDecodingForAnimatedImages.

platform/graphics/BitmapImage.h:
platform/graphics/CrossfadeGeneratedImage.cpp:

(WebCore::CrossfadeGeneratedImage::draw): Add a new argument of type DecodingMode.

platform/graphics/CrossfadeGeneratedImage.h:
platform/graphics/DecodingOptions.h: Added.

(WebCore::DecodingOptions::DecodingOptions): Default constructor: Synchronous mode.
(WebCore::DecodingOptions::operator==): Compares two DecodingOptions for equality.
(WebCore::DecodingOptions::isSynchronous): Is the frame decoded synchronously?
(WebCore::DecodingOptions::isAsynchronous): Is the frame decoded asynchronously?
(WebCore::DecodingOptions::isAsynchronousCompatibleWith): Is this DecodingOptions compatible with another one?
(WebCore::DecodingOptions::hasFullSize): Is the decoding mode asynchronous but for the image full size?
(WebCore::DecodingOptions::hasSizeForDrawing): Is this decoding mode asynchronous but for a sizeForDrawing?
(WebCore::DecodingOptions::sizeForDrawing): Returns the sizeForDrawing. m_decodingModeOrSize has to hold an IntSize.
(WebCore::DecodingOptions::maxDimension): Moved form ImageFrame.cpp.
(WebCore::DecodingOptions::has): A helper function.
(WebCore::DecodingOptions::hasDecodingMode): Does m_decodingModeOrSize a DecodingMode?
(WebCore::DecodingOptions::hasSize): Does m_decodingModeOrSize an IntSize?

platform/graphics/GeneratedImage.h: Add a new argument of type DecodingMode.
platform/graphics/GradientImage.cpp:

(WebCore::GradientImage::draw): Ditto.

platform/graphics/GradientImage.h: Ditto.
platform/graphics/GraphicsContext.cpp:

(WebCore::GraphicsContext::drawImage): Pass the ImagePaintingOptions::m_DecodingMode to Image::draw().

platform/graphics/GraphicsContext.h:

(WebCore::ImagePaintingOptions::ImagePaintingOptions): Add a new member of type DecodingMode to ImagePaintingOptions.

platform/graphics/Image.cpp:

(WebCore::Image::drawTiled): Pass DecodingMode::Synchronous to Image::draw().

platform/graphics/Image.h: Add a new argument of type DecodingMode to Image::draw().
platform/graphics/ImageFrame.cpp:

(WebCore::ImageFrame::operator=): Replace m_sizeForDrawing by m_decodingOptions.
(WebCore::ImageFrame::hasNativeImage): Check if m_nativeImage is valid and the subsamplingLevels match.
(WebCore::ImageFrame::hasFullSizeNativeImage): Checks hasNativeImage() and whether the image frame was
decoded for the image full size.
(WebCore::ImageFrame::hasDecodedNativeImageCompatibleWithOptions): Checks hasNativeImage() and the DecodingOptions match.
(WebCore::maxDimension): Deleted. Moved to DecodingOptions.h.
(WebCore::ImageFrame::isBeingDecoded): Deleted. The check for having an ImageFrame being decoded is
moved to ImageFrameCache.
(WebCore::ImageFrame::hasValidNativeImage): Deleted. No need to this function.

platform/graphics/ImageFrame.h:

(WebCore::ImageFrame::hasNativeImage): Add an std::optional<SubsamplingLevel> argument.
(WebCore::ImageFrame::hasFullSizeNativeImage): Checks whether the ImageFrame was decoded for the image full size.
(WebCore::ImageFrame::enqueueSizeForDecoding): Deleted.
(WebCore::ImageFrame::dequeueSizeForDecoding): Deleted.
(WebCore::ImageFrame::clearSizeForDecoding): Deleted.
(WebCore::ImageFrame::isBeingDecoded): Deleted.
(WebCore::ImageFrame::sizeForDrawing): Deleted.
(WebCore::ImageFrame::hasDecodedNativeImage): Deleted.
The logic of knowing whether an ImageFrame is being decoded is moved to ImageFrameCache.

platform/graphics/ImageFrameCache.cpp:

(WebCore::ImageFrameCache::cacheFrameMetadataAtIndex): Caches the metadata of an ImageFrame. If the NativeImage
was decoded for a sizeForDrawing, the size of the ImageFrame will be the nativeImageSize(). Otherwise, the
frameSizeAtIndex() will be called.
(WebCore::ImageFrameCache::cacheFrameNativeImageAtIndex): Cache a new NativeImage which involves caching new
metadata and updating the memory cache. No need to check if the existing native image is valid or not for the
DecodingOptions. Actually it would be a bug if it happens. This is why cacheNativeImageForFrameRequest() asserts
!frame.hasAsyncNativeImage() before calling cacheFrameNativeImageAtIndex().
(WebCore::ImageFrameCache::cacheAsyncFrameNativeImageAtIndex): Cache new NativeImage which was decoded asynchronously.
(WebCore::ImageFrameCache::startAsyncDecodingQueue): Call cacheAsyncFrameNativeImageAtIndex() instead of
cacheNativeImageForFrameRequest() for clarity.
(WebCore::ImageFrameCache::requestFrameAsyncDecodingAtIndex): Call hasAsyncNativeImage() instead of hasValidNativeImage()
Call frameIsDecodingCompatibleWithOptionsAtIndex() instead of frame.isBeingDecoded(). Replace the call to enqueueSizeForDecoding()
by appending the same ImageFrameRequest to m_frameCommitQueue.
(WebCore::ImageFrameCache::isAsyncDecodingQueueIdle): Use m_frameCommitQueue to answer the question whether the decodingQueue is idle.
(WebCore::ImageFrameCache::stopAsyncDecodingQueue): Use m_frameCommitQueue to loop through all the ImageFrames which are currently being decoded.
(WebCore::ImageFrameCache::frameAtIndexCacheIfNeeded): For getting the metadata, this function needs a valid frame. If it is requested
to decode the nativeImage, it has to do it synchronously.
(WebCore::ImageFrameCache::singlePixelSolidColor): Don't cache the frame if it is an animated image or the size is not a single pixel.
(WebCore::ImageFrameCache::frameIsBeingDecodedAndIsCompatibleWithOptionsAtIndex): Use m_frameCommitQueue to answer the question whether an ImageFrame
is being decoded and is compatible with DecodingOptions.
(WebCore::ImageFrameCache::frameHasFullSizeNativeImageAtIndex): Calls ImageFrame::hasFullNativeImage() for a frame.
(WebCore::ImageFrameCache::frameHasDecodedNativeImageCompatibleWithOptionsAtIndex): Calls ImageFrame::hasDecodedNativeImageCompatibleWithOptions() for a frame.
(WebCore::ImageFrameCache::frameImageAtIndex): Returns the current NativeImage without caching.
(WebCore::ImageFrameCache::frameImageAtIndexCacheIfNeeded): Returns the current NativeImage but cache it synchronously if needed.
(WebCore::ImageFrameCache::setFrameNativeImageAtIndex): Deleted.
(WebCore::ImageFrameCache::setFrameMetadataAtIndex): Deleted.
(WebCore::ImageFrameCache::replaceFrameNativeImageAtIndex): Deleted.
(WebCore::ImageFrameCache::frameIsBeingDecodedAtIndex): Deleted.
(WebCore::ImageFrameCache::frameHasImageAtIndex): Deleted.
(WebCore::ImageFrameCache::frameHasValidNativeImageAtIndex): Deleted.
(WebCore::ImageFrameCache::frameHasDecodedNativeImage): Deleted.

platform/graphics/ImageFrameCache.h: Two ImageFrameRequest queues will be used.

-- The existing one m_frameRequestQueue which is shared between the main thread and decoding thread. The requests will be
dequeued from it before starting the decoding. The decoded NativeImage will be cached only on the main thread. The decoding
thread is not blocked by the callOnMainThread(). This means there might be multiple ImageFrameRequests which were dequeued
while their NativeImages have not been cached yet.
-- A new one m_frameCommitQueue which is track all the ImageFrameRequests whose NativeImages have not been cached yet.
(WebCore::ImageFrameCache::frameAtIndexCacheIfNeeded): Be explicit about caching the image frame. frameImageAtIndex()
returns the current image frame without caching. frameAtIndexCacheIfNeeded(). returns the current image frame but cache
it if needed.
(WebCore::ImageFrameCache::ImageFrameRequest::operator==): Compares two ImageFrameRequests for equality.

platform/graphics/ImageSource.cpp:

(WebCore::ImageSource::frameImageAtIndexCacheIfNeeded):
(WebCore::ImageSource::frameImageAtIndex): Deleted.

platform/graphics/ImageSource.h:

(WebCore::ImageSource::requestFrameAsyncDecodingAtIndex): Change the type of the argument from IntSize to be const std::optional<IntSize>.
(WebCore::ImageSource::frameIsBeingDecodedAndIsCompatibleWithOptionsAtIndex): Rename of frameIsBeingDecodedAtIndex(). Replace the argument of type
std::optional<IntSize> by an argument of type DecodingOptions.
(WebCore::ImageSource::frameHasFullSizeNativeImageAtIndex): A wrapper around the ImageFrameCache function.
(WebCore::ImageSource::frameHasDecodedNativeImageCompatibleWithOptionsAtIndex): Ditto.
(WebCore::ImageSource::frameImageAtIndex): Ditto.
(WebCore::ImageSource::frameIsBeingDecodedAtIndex): Deleted.
(WebCore::ImageSource::frameHasValidNativeImageAtIndex): Deleted.
(WebCore::ImageSource::frameHasDecodedNativeImage): Deleted.

platform/graphics/NamedImageGeneratedImage.cpp:

(WebCore::NamedImageGeneratedImage::draw): Add a new argument of type DecodingMode.

platform/graphics/NamedImageGeneratedImage.h: Ditto.
platform/graphics/cairo/ImageBufferCairo.cpp:

(WebCore::ImageBuffer::draw): Add a new argument of type DecodingMode.

platform/graphics/cg/ImageDecoderCG.cpp:

(WebCore::ImageDecoder::createFrameImageAtIndex): Replace the sizeForDrawing argument by a DecodingMode argument. Add a new handling
for decoding asynchronously for the image full size.

platform/graphics/cg/ImageDecoderCG.h: Change the prototype of the function.
platform/graphics/cg/PDFDocumentImage.cpp:

(WebCore::PDFDocumentImage::draw): Add a new argument of type DecodingMode.

platform/graphics/cg/PDFDocumentImage.h:
platform/graphics/win/ImageCGWin.cpp:

(WebCore::BitmapImage::getHBITMAPOfSize): Pass DecodingMode::Synchronous to Image::draw().
(WebCore::BitmapImage::drawFrameMatchingSourceSize): Ditto.

platform/graphics/win/ImageDecoderDirect2D.cpp:

(WebCore::ImageDecoder::createFrameImageAtIndex): Replace the sizeForDrawing argument by a DecodingMode argument.

platform/graphics/win/ImageDecoderDirect2D.h: Change the prototype of the function.
platform/image-decoders/ImageDecoder.cpp:

(WebCore::ImageDecoder::createFrameImageAtIndex): Replace the sizeForDrawing argument by a DecodingMode argument.

platform/image-decoders/ImageDecoder.h: Change the prototype of the function.
rendering/RenderBoxModelObject.cpp:

(WebCore::RenderBoxModelObject::paintFillLayerExtended): Draw the background image asynchronously if the image size is large.

rendering/RenderImage.cpp:

(WebCore::RenderImage::paintIntoRect): Draw the background image element asynchronously if the image size is large.

svg/graphics/SVGImage.cpp:

(WebCore::SVGImage::drawForContainer): Pass DecodingMode::Synchronous to draw().
(WebCore::SVGImage::nativeImageForCurrentFrame): Ditto.
(WebCore::SVGImage::nativeImage): Ditto.
(WebCore::SVGImage::draw): Add a new argument of type DecodingMode.

svg/graphics/SVGImage.h: Change the prototype of the function.
svg/graphics/SVGImageForContainer.cpp:

(WebCore::SVGImageForContainer::draw): Add a new argument of type DecodingMode.

svg/graphics/SVGImageForContainer.h: Change the prototype of the function.

LayoutTests:

fast/images/animated-image-different-dest-size-expected.html: Added.
fast/images/animated-image-different-dest-size.html: Added.

This test crashes without this patch.

fast/images/animated-image-loop-count.html:

Clear the memory cache so the test can be not flaky. Running it with -repeat-each was failing.

fast/images/async-image-background-image-expected.html: Added.
fast/images/async-image-background-image.html: Added.

Ensures the background image can be drawn asynchronously if it is large.

fast/images/async-image-canvas-draw-image-expected.html: Added.
fast/images/async-image-canvas-draw-image.html: Added.

Ensures the image is drawn synchronously on the canvas regardless of its size.

fast/images/ordered-animated-image-frames.html:

Clear the memory cache so the test can be not flaky. Running it with -repeat-each was failing.

fast/images/reset-image-animation-expected.txt:
fast/images/reset-image-animation.html:

Change how the steps of the test are ordered so the test can be not flaky.
Running it with -repeat-each was failing.

fast/images/resources/red-green-blue-900-300.png: Added.

7:06 PM Changeset in webkit [214449] by aestes@apple.com

3 edits in trunk/Source/WebKit2

[iOS] Suppress page count recalc when setting a snapshot paper rect
https://bugs.webkit.org/show_bug.cgi?id=170156
<rdar://problem/30787872>

Reviewed by Tim Horton.

When _WKWebViewPrintFormatter computes a page count for snapshotting, it updates the paper
rect, which triggers UIKit to ask for the page count again while we're in the middle of
printing, putting us in a bad state.

Suppress page count recalc while setting the new paper rect, since we know this will not
cause it to change.

UIProcess/_WKWebViewPrintFormatter.mm:

(-[_WKWebViewPrintFormatter _setSnapshotPaperRect:]): Temporarily set
_suppressPageCountRecalc to YES before updating the paper rect.
(-[_WKWebViewPrintFormatter _setNeedsRecalc]): Only call [super _setNeedsRecalc] if
_suppressPageCountRecalc is NO.

UIProcess/_WKWebViewPrintFormatterInternal.h: Declared -_setNeedsRecalc.

7:02 PM Changeset in webkit [214448] by Wenson Hsieh

2 edits in trunk/Source/WebKit2

[WK2] Callout Bar should be dismissed during data interaction
https://bugs.webkit.org/show_bug.cgi?id=170109
<rdar://problem/31171570>

Reviewed by Tim Horton.

Minor tweak to get the callout bar to duck during data interaction. Import a new UIKit SPI header in WebKit2.

Platform/spi/ios/UIKitSPI.h:

6:48 PM Changeset in webkit [214447] by Lucas Forschler

1 edit in trunk/Tools/ChangeLog

Test commit against new svn server.

6:31 PM Changeset in webkit [214446] by Lucas Forschler

2 edits in trunk/Tools

<rdar://problem/30949128>
Update build.webkit.org to use port 16000 for PBChangeSource.

Reviewed by Alexey Proskuryakov.

BuildSlaveSupport/build.webkit.org-config/master.cfg:

5:32 PM Changeset in webkit [214445] by commit-queue@webkit.org

2 edits in trunk/Source/WebCore

Activate release libwebrtc logging when WebRTC log channel is on
https://bugs.webkit.org/show_bug.cgi?id=169659

Patch by Youenn Fablet <youenn@apple.com> on 2017-03-27
Reviewed by Alex Christensen.

platform/mediastream/libwebrtc/LibWebRTCProvider.cpp:

(WebCore::initializePeerConnectionFactoryAndThreads):

5:16 PM Changeset in webkit [214444] by Keith Rollin

2 edits in trunk/Source/WebKit2

Update NetworkResourceLoader logging statement
https://bugs.webkit.org/show_bug.cgi?id=170126

Reviewed by Alex Christensen.

NetworkResourceLoader contains a logging statement that prints, in
part: "start: Retrieving resource from cache". This is misleading,
since it's not necessarily the case that the resource will be loaded
from the cache. What is actually happening at the point where that is
logged is that the cache is being checked for the resource. It still
remains to be seen if the resource will be found in it. To address
this, change the logging to say "start: Checking cache for resource".

NetworkProcess/NetworkResourceLoader.cpp:

(WebKit::NetworkResourceLoader::start):

5:06 PM Changeset in webkit [214443] by Antti Koivisto

10 edits in trunk/Source/WebCore

Move visibleInViewportStateChanged callback from Element to render tree
https://bugs.webkit.org/show_bug.cgi?id=170039

Reviewed by Zalan Bujtas.

Make it easier to use from the render tree.

Also for simplicity move the bits from RenderObject rare data to RenderElement.
There is plenty of space there.

dom/Element.h:

(WebCore::Element::isVisibleInViewportChanged): Deleted.

html/HTMLMediaElement.h:
rendering/RenderElement.cpp:

(WebCore::RenderElement::RenderElement):
(WebCore::RenderElement::willBeDestroyed):
(WebCore::RenderElement::registerForVisibleInViewportCallback):
(WebCore::RenderElement::unregisterForVisibleInViewportCallback):
(WebCore::RenderElement::setVisibleInViewportState):
(WebCore::RenderElement::visibleInViewportStateChanged):

rendering/RenderElement.h:

(WebCore::RenderElement::visibleInViewportState):

rendering/RenderObject.cpp:

(WebCore::RenderObject::setIsRegisteredForVisibleInViewportCallback): Deleted.
(WebCore::RenderObject::setVisibleInViewportState): Deleted.

rendering/RenderObject.h:

(WebCore::RenderObject::hasOutlineAutoAncestor):
(WebCore::RenderObject::RenderObjectRareData::RenderObjectRareData):
(WebCore::RenderObject::isRegisteredForVisibleInViewportCallback): Deleted.
(WebCore::RenderObject::visibleInViewportState): Deleted.

rendering/RenderVideo.cpp:

(WebCore::RenderVideo::visibleInViewportStateChanged):

rendering/RenderVideo.h:
rendering/RenderView.cpp:

(WebCore::RenderView::updateVisibleViewportRect):

4:37 PM Changeset in webkit [214442] by commit-queue@webkit.org

5 edits in trunk/Source

Web Inspector: Saving files should not suggest the top level directory
https://bugs.webkit.org/show_bug.cgi?id=170085
<rdar://problem/30197794>

Patch by Joseph Pecoraro <Joseph Pecoraro> on 2017-03-27
Reviewed by Matt Baker.

Source/WebKit/mac:

WebCoreSupport/WebInspectorClient.mm:

(WebInspectorFrontendClient::save):
Don't set panel.directoryURL unless we have a good suggestion.

Source/WebKit2:

UIProcess/mac/RemoteWebInspectorProxyMac.mm:

(WebKit::RemoteWebInspectorProxy::platformSave):

UIProcess/mac/WebInspectorProxyMac.mm:

(WebKit::WebInspectorProxy::platformSave):
Don't set panel.directoryURL unless we have a good suggestion.

4:36 PM Changeset in webkit [214441] by commit-queue@webkit.org

13 edits in trunk

addIceCandidate should not throw if passed null or undefined
https://bugs.webkit.org/show_bug.cgi?id=170118

Patch by Youenn Fablet <youenn@apple.com> on 2017-03-27
Reviewed by Eric Carlson.

LayoutTests/imported/w3c:

web-platform-tests/webrtc/rtcpeerconnection/rtcpeerconnection-idl-expected.txt:

Source/WebCore:

Covered by updated test.

A null/undefined candidate passed to addIceCandidate means end of Ice candidate..

Modules/mediastream/PeerConnectionBackend.cpp:

(WebCore::PeerConnectionBackend::addIceCandidate):

Modules/mediastream/PeerConnectionBackend.h:

(WebCore::PeerConnectionBackend::endOfIceCandidates):

Modules/mediastream/RTCPeerConnection.cpp:

(WebCore::RTCPeerConnection::queuedAddIceCandidate):

Modules/mediastream/RTCPeerConnection.h:
Modules/mediastream/RTCPeerConnection.idl:
Modules/mediastream/RTCPeerConnection.js:

(addIceCandidate):

LayoutTests:

Updating test to log addIceCandidate rejection.

webrtc/datachannel/basic.html:
webrtc/routines.js:

(iceCallback1):
(iceCallback2):
(onAddIceCandidateError):

4:32 PM Changeset in webkit [214440] by commit-queue@webkit.org

3 edits in trunk/Source/WebKit2

Add delegate method to let an WebKit client provide a custom image preview controller.
https://bugs.webkit.org/show_bug.cgi?id=170076
rdar://problem/31153051

Add a new delegate method (_webView:previewViewControllerForImage:alternateURL:defaultActions:elementInfo:) to
WKUIDelegatePrivate, an client can use this method to provide a custom image preview view controller if the image
has alternated URL or custom image info.

Patch by Yongjun Zhang <yongjun_zhang@apple.com> on 2017-03-27
Reviewed by Beth Dakin.

UIProcess/API/Cocoa/WKUIDelegatePrivate.h: Add delegate method _webView:previewViewControllerForImage:alternateURL:defaultActions:elementInfo:.
UIProcess/ios/WKContentViewInteraction.mm:

(-[WKContentView _presentedViewControllerForPreviewItemController:]): If the image has alternated URL or custom info, try to use the client-provided

preview view controller. Otherwise, use the default WKImagePreviewViewController.

4:31 PM Changeset in webkit [214439] by webkit@devinrousso.com

12 edits in trunk/Source/WebInspectorUI

Web Inspector: The inspector sidebar has weird sub-column sizing issues
https://bugs.webkit.org/show_bug.cgi?id=161966

Reviewed by Joseph Pecoraro.

UserInterface/Views/Sidebar.js:

(WebInspector.Sidebar.prototype.resizerDragEnded):
(WebInspector.Sidebar.prototype._recalculateWidth):
Ensure that the selected sidebar panel is notified of the size change (sizeDidChange) when
the resizer drag ends.

UserInterface/Views/DetailsSectionDataGridRow.js:

(WebInspector.DetailsSectionDataGridRow.prototype.sizeDidChange):

UserInterface/Views/ProbeSetDetailsSection.js:

(WebInspector.ProbeSetDetailsSection.prototype.sizeDidChange):
Add/update the sizeDidChange protected function.

UserInterface/Views/ComputedStyleDetailsPanel.js:

(WebInspector.ComputedStyleDetailsPanel.prototype.initialLayout):
(WebInspector.ComputedStyleDetailsPanel.prototype.sizeDidChange):

UserInterface/Views/DOMNodeDetailsSidebarPanel.js:

(WebInspector.DOMNodeDetailsSidebarPanel.prototype.sizeDidChange):

UserInterface/Views/LayerTreeDetailsSidebarPanel.js:

(WebInspector.LayerTreeDetailsSidebarPanel.prototype.sizeDidChange):

UserInterface/Views/ResourceDetailsSidebarPanel.js:

(WebInspector.ResourceDetailsSidebarPanel.prototype.sizeDidChange):

UserInterface/Views/ProbeDetailsSidebarPanel.js:

(WebInspector.ProbeDetailsSidebarPanel.prototype.sizeDidChange):
Update the contained DataGrid elements whenever the sidebar size changes so that the resizer
elements are positioned correctly.

UserInterface/Views/View.js:

(WebInspector.View.prototype.updateLayoutIfNeeded):

UserInterface/Views/TimelineOverview.js:

(WebInspector.TimelineOverview.prototype.updateLayoutIfNeeded):

UserInterface/Views/TimelineRuler.js:

(WebInspector.TimelineRuler.prototype.updateLayoutIfNeeded):
Ensure the layoutReason is also propagated.

4:24 PM Changeset in webkit [214438] by jfbastien@apple.com

4 edits
2 adds in trunk

WebAssembly: misc memory testing
https://bugs.webkit.org/show_bug.cgi?id=170137

Reviewed by Keith Miller.

JSTests:

wasm/assert.js: handle newlines in code we print out, avoid regex
wasm/function-tests/memory-import-and-grow.js: Added.

(const.instantiate):
(const.test):

wasm/function-tests/memory-section-and-import.js: Added.

(const.instantiate):

Source/JavaScriptCore:

wasm/js/WebAssemblyInstanceConstructor.cpp:

(JSC::WebAssemblyInstanceConstructor::createInstance): improve error messages

4:21 PM Changeset in webkit [214437] by Ryan Haddad

1 edit
3 adds in trunk/LayoutTests

Rebaseline svg/css/getComputedStyle-basic.xhtml for macOS.

Unreviewed test gardening.

platform/mac-elcapitan/svg/css/getComputedStyle-basic-expected.txt: Added.
platform/mac/svg/css/getComputedStyle-basic-expected.txt: Added.

4:19 PM Changeset in webkit [214436] by commit-queue@webkit.org

3 edits in trunk/Source/WebInspectorUI

Web Inspector: Resource Details Sidebar displays previous image metrics when viewing resource where content load failed
https://bugs.webkit.org/show_bug.cgi?id=170065
<rdar://problem/27081591>

Patch by Joseph Pecoraro <Joseph Pecoraro> on 2017-03-27
Reviewed by Matt Baker.

UserInterface/Models/Resource.js:

(WebInspector.Resource.prototype.getImageSize.imageDidLoad):
(WebInspector.Resource.prototype.getImageSize.requestContentFailure):
(WebInspector.Resource.prototype.getImageSize):
If the requestContent failed, return a null image size.

UserInterface/Views/ResourceDetailsSidebarPanel.js:

(WebInspector.ResourceDetailsSidebarPanel.prototype._refreshImageSizeSection.hideImageSection):
(WebInspector.ResourceDetailsSidebarPanel.prototype._refreshImageSizeSection):
If the callback is called with a null size, hide the image size section.

3:56 PM Changeset in webkit [214435] by Antti Koivisto

15 edits
2 adds in trunk

Source/WebCore:
Allow the page to render before <link> stylesheet tags in body
https://bugs.webkit.org/show_bug.cgi?id=149157
<rdar://problem/24658830>

Reviewed by Simon Fraser.

Currently we block style and renderer building completely if document has any loading
stylesheets. In case a script queries something layout dependent we construct the render
tree with whatever style we have but block painting in it.

This patch changes behavior so that a loading stylesheet in body only blocks rendering for elements
that are after it. The expectation is that such stylesheets rarely affect elements before them
and the elements can be rendered without causing ugly visible styling changes.

The patch replaces the old flash-of-unstyled-content (FOUC) preventation mechanism with a more
fine-grained one. Paint blocking is now done on per-renderer basis with based on isNonFinal flag in
RenderStyle.

For stylesheets in head the behavior should be largely unchanged.

Test: http/tests/incremental/stylesheet-body-incremental-rendering.html

css/StyleResolver.cpp:

(WebCore::StyleResolver::pseudoStyleRulesForElement):

dom/Document.cpp:

(WebCore::Document::Document):
(WebCore::Document::resolveStyle):
(WebCore::Document::updateLayoutIgnorePendingStylesheets):

Remove the old FOUC preventation state tracking.

(WebCore::Document::shouldScheduleLayout):
(WebCore::Document::didRemoveAllPendingStylesheet):

Repaints will now get triggered by the normal style mechanism.

dom/Document.h:

(WebCore::Document::hasNodesWithNonFinalStyle):
(WebCore::Document::setHasNodesWithNonFinalStyle):

Track if we need to recompute the style later because non-final or unstyled elements.

(WebCore::Document::didLayoutWithPendingStylesheets): Deleted.
(WebCore::Document::hasNodesWithPlaceholderStyle): Deleted.
(WebCore::Document::setHasNodesWithPlaceholderStyle): Deleted.

html/HTMLFrameSetElement.cpp:

(WebCore::HTMLFrameSetElement::rendererIsNeeded):

page/FrameView.cpp:

(WebCore::FrameView::qualifiesAsVisuallyNonEmpty):

Don't qualify as visually non-empty if we have loading stylesheets in head (even if there is
a fouc-prevented render tree).

(WebCore::FrameView::fireLayoutRelatedMilestonesIfNeeded):

rendering/RenderBlock.cpp:

(WebCore::RenderBlock::paintContents):

Instead of a global test, block painting if isNonFinal is set in the renderer's style.

rendering/RenderLayer.cpp:

(WebCore::shouldSuppressPaintingLayer):

rendering/style/RenderStyle.cpp:

(WebCore::RenderStyle::changeRequiresRepaint):

The isNonFinal flag prevents painting so we need to trigger repaint when it gets cleared.

rendering/style/RenderStyle.h:

(WebCore::RenderStyle::isNotFinal):
(WebCore::RenderStyle::setIsNotFinal):
(WebCore::RenderStyle::isPlaceholderStyle): Deleted.
(WebCore::RenderStyle::setIsPlaceholderStyle): Deleted.

There is no need for placeholder styles anymore. Reuse the bit for isNotFinal.

rendering/style/StyleRareNonInheritedData.cpp:

(WebCore::StyleRareNonInheritedData::StyleRareNonInheritedData):
(WebCore::StyleRareNonInheritedData::operator==):

rendering/style/StyleRareNonInheritedData.h:
style/StyleScope.cpp:

(WebCore::Style::Scope::analyzeStyleSheetChange):
(WebCore::Style::Scope::updateActiveStyleSheets):

style/StyleTreeResolver.cpp:

(WebCore::Style::TreeResolver::styleForElement):
(WebCore::Style::TreeResolver::resolveElement):

If we have seens a loading stylesheet and don't have a renderer yet don't style the element.
In case there is a renderer or we are ignoring pending sheets, resolve the style normally
but mark it as non-final.

(WebCore::Style::makePlaceholderStyle): Deleted.

LayoutTests:
Loading in-body stylesheets should not block rendering of elements before them
https://bugs.webkit.org/show_bug.cgi?id=169345

Reviewed by Simon Fraser.

http/tests/incremental/resources/delayed-css.php: Added.
http/tests/incremental/stylesheet-body-incremental-rendering-expected.html: Added.
http/tests/incremental/stylesheet-body-incremental-rendering.html: Added.

3:22 PM Changeset in webkit [214434] by jmarcell@apple.com

4 edits in trunk/Tools

Bots should run the dashboard tests
https://bugs.webkit.org/show_bug.cgi?id=168994

Reviewed by Daniel Bates and David Kilzer.

We pull the --results-directory argument value out so that RunDashboardTests can override it so that we can separately run the dashboard layout tests.

BuildSlaveSupport/build.webkit.org-config/master.cfg:

(RunWebKitTests): Define a class variable called resultDirectory for the --results-directory argument so that subclasses can override this if needed.
(RunWebKitTests.start): Set the --results-directory argument value to the value of self.resultDirectory.
(RunDashboardTests): Subclass of RunWebKitTests in which we run the layout tests for the Bot Watcher's Dashboard.
(RunDashboardTests.start): Override the --layout-tests-directory to point to the dashboard layout tests.
(ExtractTestResults.addCustomURLs): Update the current results URL label to indicate that it links to layout tests results and add a new link to link to dashboard layout test results.
(TestFactory.init): Update unit tests to account for new RunDashboardTests step.

BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:
BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/BuildbotIteration.js:

3:09 PM Changeset in webkit [214433] by mmaxfield@apple.com

5 edits
2 adds in trunk

Test variation font ranges in the CSS Font Loading API
https://bugs.webkit.org/show_bug.cgi?id=170022

Reviewed by Dean Jackson.

Source/WebCore:

Test: fast/text/variations/font-loading-api-parse-ranges.html

Don't use keywords when reporting font variation range values.

css/CSSComputedStyleDeclaration.cpp:

(WebCore::ComputedStyleExtractor::fontNonKeywordWeightFromStyleValue):
(WebCore::ComputedStyleExtractor::fontWeightFromStyleValue):
(WebCore::ComputedStyleExtractor::fontNonKeywordStretchFromStyleValue):
(WebCore::ComputedStyleExtractor::fontStretchFromStyleValue):
(WebCore::ComputedStyleExtractor::fontNonKeywordStyleFromStyleValue):
(WebCore::ComputedStyleExtractor::fontStyleFromStyleValue):

css/CSSComputedStyleDeclaration.h:
css/FontFace.cpp:

(WebCore::FontFace::style):
(WebCore::FontFace::weight):
(WebCore::FontFace::stretch):

LayoutTests:

fast/text/variations/font-loading-api-parse-ranges-expected.txt: Added.
fast/text/variations/font-loading-api-parse-ranges.html: Added.

2:49 PM Changeset in webkit [214432] by Jonathan Bedard

3 edits in trunk/Tools

webkitpy: Look for 'ios' test expectations for IOSPorts
https://bugs.webkit.org/show_bug.cgi?id=169413

Reviewed by Alexey Proskuryakov.

Test expectations for iOS tests should include the 'ios' folder. Modify
default_baseline_search_path to include 'ios' folder and derivatives for
IOSPort.

Scripts/webkitpy/port/ios.py:

(IOSPort._generate_all_test_configurations): Moved from IOSSimulatorPort.
(IOSPort.default_baseline_search_path): Add ios expectations along with port specific expectations.
(IOSPort.test_expectations_file_position): Added for testing.

Scripts/webkitpy/port/ios_simulator.py:

(IOSSimulatorPort._generate_all_test_configurations): Moved to IOSPort.
(IOSSimulatorPort.default_baseline_search_path): Moved to IOSPort.

2:42 PM Changeset in webkit [214431] by Jonathan Bedard

2 edits in trunk/Tools

REGRESSION: Reseting _device_map on each child process
<rdar://problem/31274476>

Unreviewed infrastructure fix.

Scripts/webkitpy/port/ios_simulator.py:

(IOSSimulatorPort): Add _DEVICE_MAP class variable.
(IOSSimulatorPort.init): Use shared class variable
(IOSSimulatorPort._device_for_worker_number_map): Use _DEVICE_MAP class variable
(IOSSimulatorPort._create_devices): Ditto.
(IOSSimulatorPort._quit_ios_simulator): Ditto.
(IOSSimulatorPort.clean_up_test_run): Ditto.

2:32 PM Changeset in webkit [214430] by andersca@apple.com

2 edits in trunk/Source/WebCore

Propagate shipping contact update errors
https://bugs.webkit.org/show_bug.cgi?id=170141
rdar://problem/31276576

Reviewed by Tim Horton.

Modules/applepay/ApplePaySession.cpp:

(WebCore::convertAndValidate):

2:20 PM Changeset in webkit [214429] by Ryan Haddad

11 edits
15 deletes in trunk

Unreviewed, rolling out r214411.

Two of the LayoutTests for this change time out on ios-
simulator.

Reverted changeset:

"[Modern Media Controls] Improve media documents across macOS,
iPhone and iPad"
https://bugs.webkit.org/show_bug.cgi?id=169145
http://trac.webkit.org/changeset/214411

1:55 PM Changeset in webkit [214428] by andersca@apple.com

2 edits in trunk/Source/WebKit2

Fix uses of deprecated API
https://bugs.webkit.org/show_bug.cgi?id=170138
rdar://problem/28882449

Reviewed by Tim Horton.

UIProcess/ApplePay/cocoa/WebPaymentCoordinatorProxyCocoa.mm:

(WebKit::toPKContactFields):
(WebKit::toPKAddressField):
(WebKit::toPKPaymentRequest):
Don't use deprecated APIs.

(WebKit::toNSError):
Don't try to insert nil in a dictionary.

12:57 PM Changeset in webkit [214427] by achristensen@apple.com

4 edits in trunk/Source/ThirdParty/libwebrtc

Build libwebrtc with even more warnings
https://bugs.webkit.org/show_bug.cgi?id=169997

Reviewed by Tim Horton.

There are still OSAtomic* functions I don't want to worry about right now,
so I'm keeping a few -Wno-deprecated-declarations, but everything else can go.

Configurations/libsrtp.xcconfig:
Configurations/libwebrtc.xcconfig:
Configurations/libwebrtcpcrtc.xcconfig:

12:44 PM Changeset in webkit [214426] by graouts@webkit.org

5 edits in trunk

[Modern Media Controls] Clicking on the tracks button when the tracks panel is up in a media document pauses the video
https://bugs.webkit.org/show_bug.cgi?id=168517
<rdar://problem/30577636>

Reviewed by Dean Jackson.

Source/WebCore:

We completely turn off default event handling in MediaDocument.cpp since we're implementing the
behavior we expect to pause and resume the video in the modern-media-controls module already. This
gets rid of this odd case where the content would not see the "click" event while the C++ side would
handle it and pause the video.

Modules/modern-media-controls/media/media-controller.js:

(MediaController):
(MediaController.prototype.handleEvent):
(MediaController.prototype._containerWasClicked): Deleted.

html/MediaDocument.cpp:

(WebCore::MediaDocument::defaultEventHandler):

LayoutTests:

media/video-click-dblckick-standalone.html: We disable modern-media-controls here since we know that this test won't pass with them on.

12:40 PM Changeset in webkit [214425] by Ryan Haddad

2 edits in trunk/LayoutTests

Mark media/modern-media-controls/time-label/time-label-white-space-nowrap.html as flaky.
https://bugs.webkit.org/show_bug.cgi?id=169679

Unreviewed test gardening.

platform/mac/TestExpectations:

12:36 PM Changeset in webkit [214424] by Ryan Haddad

2 edits in trunk/LayoutTests

Mark inspector/dom-debugger/dom-breakpoints.html as flaky.
https://bugs.webkit.org/show_bug.cgi?id=170127

Unreviewed test gardening.

platform/mac/TestExpectations:

12:16 PM Changeset in webkit [214423] by Ryan Haddad

2 edits in trunk/LayoutTests

Mark media/modern-media-controls/time-label/time-label.html as flaky.
https://bugs.webkit.org/show_bug.cgi?id=169678

Unreviewed test gardening.

platform/mac-wk1/TestExpectations:

12:09 PM Changeset in webkit [214422] by Ryan Haddad

2 edits in trunk/LayoutTests

Marking two modern-media-controls tests as timeouts.
https://bugs.webkit.org/show_bug.cgi?id=170123

Unreviewed test gardening.

platform/ios-simulator/TestExpectations:

11:13 AM Changeset in webkit [214421] by commit-queue@webkit.org

10 edits in trunk

Tighten RTCDatachannel creation and parameter getters
https://bugs.webkit.org/show_bug.cgi?id=170081

Patch by Youenn Fablet <youenn@apple.com> on 2017-03-27
Reviewed by Eric Carlson.

Source/WebCore:

Covered by updated tests.

Adding some parameter checks when creating data channels.
Making some getters nullable as per the spec.

Modules/mediastream/RTCDataChannel.h:
Modules/mediastream/RTCDataChannel.idl:
Modules/mediastream/RTCPeerConnection.cpp:

(WebCore::RTCPeerConnection::createDataChannel):

Modules/mediastream/RTCPeerConnection.idl:
Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.cpp:

(WebCore::LibWebRTCMediaEndpoint::createDataChannel):
(WebCore::LibWebRTCMediaEndpoint::addDataChannel):

platform/mediastream/RTCDataChannelHandler.h:

(): Deleted.

LayoutTests:

webrtc/datachannel/basic-expected.txt:
webrtc/datachannel/basic.html:

11:10 AM Changeset in webkit [214420] by commit-queue@webkit.org

23 edits
4 adds in trunk

Add support for RTCRtpReceiver/RTCRtpSender getParameters
https://bugs.webkit.org/show_bug.cgi?id=170057

Patch by Youenn Fablet <youenn@apple.com> on 2017-03-27
Reviewed by Alex Christensen.

Source/ThirdParty/libwebrtc:

Source/webrtc/api/mediatypes.h:

Source/WebCore:

Test: webrtc/video-getParameters.html

getParameters returns a dictionary of values taken from libwebrtc RtpReceiverInterface/RtpSenderInrterface objects.
Added a direct link between WebCore RTCRtpReceiver and libwebrtc RtpReceiverInterface object.
Making the link between WebCore RTCRtpSender and libwebrtc RtpSenderInrterface object through
PeerConnectionBackend to keep the current architecture shared with OpenWebRTC.
In the future, we should try to make the link more direct.

Added routines to support the conversion from libwebrtc to WebCore.
Ensured that RTCRtpReceiver is cleaning its backend when being stopped.

CMakeLists.txt:
DerivedSources.make:
Modules/mediastream/MediaEndpointPeerConnection.cpp:

(WebCore::MediaEndpointPeerConnection::setRemoteDescriptionTask):

Modules/mediastream/PeerConnectionBackend.h:

(WebCore::PeerConnectionBackend::getParameters):

Modules/mediastream/RTCPeerConnection.cpp:

(WebCore::RTCPeerConnection::doClose):
(WebCore::RTCPeerConnection::getParameters):

Modules/mediastream/RTCPeerConnection.h:
Modules/mediastream/RTCRtpParameters.h: Added.
Modules/mediastream/RTCRtpParameters.idl: Added.
Modules/mediastream/RTCRtpReceiver.cpp:

(WebCore::RTCRtpReceiver::RTCRtpReceiver):

Modules/mediastream/RTCRtpReceiver.h:

(WebCore::RTCRtpReceiver::Backend::~Backend):
(WebCore::RTCRtpReceiver::Backend::getParameters):
(WebCore::RTCRtpReceiver::create):
(WebCore::RTCRtpReceiver::stop):
(WebCore::RTCRtpReceiver::setBackend):
(WebCore::RTCRtpReceiver::getParameters):

Modules/mediastream/RTCRtpReceiver.idl:
Modules/mediastream/RTCRtpSender.cpp:

(WebCore::RTCRtpSender::create):
(WebCore::RTCRtpSender::RTCRtpSender):
(WebCore::RTCRtpSender::replaceTrack):
(WebCore::RTCRtpSender::getParameters):

Modules/mediastream/RTCRtpSender.h:

(WebCore::RTCRtpSender::Backend::~Backend):
(WebCore::RTCRtpSender::isStopped):
(WebCore::RTCRtpSender::stop):

Modules/mediastream/RTCRtpSender.idl:
Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.cpp:

(WebCore::LibWebRTCMediaEndpoint::addTrack):
(WebCore::LibWebRTCMediaEndpoint::addRemoteTrack):
(WebCore::LibWebRTCMediaEndpoint::OnAddTrack):
(WebCore::LibWebRTCMediaEndpoint::stop):
(WebCore::fillEncodingParameters):
(WebCore::fillHeaderExtensionParameters):
(WebCore::fillCodecParameters):
(WebCore::fillRtpParameters):
(WebCore::RTCRtpReceiverBackend::getParameters):
(WebCore::LibWebRTCMediaEndpoint::getRTCRtpSenderParameters):

Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.h:
Modules/mediastream/libwebrtc/LibWebRTCPeerConnectionBackend.cpp:

(WebCore::LibWebRTCPeerConnectionBackend::notifyAddedTrack):
(WebCore::LibWebRTCPeerConnectionBackend::getParameters):

Modules/mediastream/libwebrtc/LibWebRTCPeerConnectionBackend.h:
WebCore.xcodeproj/project.pbxproj:
platform/mediastream/mac/RealtimeOutgoingAudioSource.h:

LayoutTests:

webrtc/video-getParameters-expected.txt: Added.
webrtc/video-getParameters.html: Added.

10:50 AM Changeset in webkit [214419] by mmaxfield@apple.com

17 edits in trunk

font variation properties don't need to accept numbers
https://bugs.webkit.org/show_bug.cgi?id=169357

Reviewed by Antti Koivisto.

Source/WebCore:

The CSS Fonts level 4 spec stabilized the grammar accepted by font-weight,
font-style, and font-stretch. The changes are that font-style and
font-stretch no longer accept raw numbers, and the @font-face descriptor
ranges are now separated by spaces instead of slashes.

Tests: fast/text/font-selection-font-face-parse.html

fast/text/font-selection-font-loading-api-parse.html
fast/text/font-stretch-parse.html
fast/text/font-style-parse.html
fast/text/variations/font-selection-properties.html

css/CSSComputedStyleDeclaration.cpp:

(WebCore::ComputedStyleExtractor::fontWeightFromStyleValue):
(WebCore::fontWeightFromStyle):
(WebCore::ComputedStyleExtractor::fontStretchFromStyleValue):
(WebCore::fontStretchFromStyle):
(WebCore::ComputedStyleExtractor::fontStyleFromStyleValue):
(WebCore::fontStyleFromStyle):

css/CSSComputedStyleDeclaration.h:
css/FontFace.cpp:

(WebCore::FontFace::style):
(WebCore::FontFace::weight):
(WebCore::FontFace::stretch):
(WebCore::rangeIsSingleValue): Deleted.

css/StyleBuilderConverter.h:

(WebCore::StyleBuilderConverter::convertFontStretchFromValue):

css/parser/CSSPropertyParser.cpp:

(WebCore::consumeFontWeightRange):
(WebCore::consumeFontStretch):
(WebCore::consumeFontStretchRange):
(WebCore::consumeFontStyle):
(WebCore::consumeFontStyleRange):

LayoutTests:

fast/text/font-selection-font-face-parse-expected.txt:
fast/text/font-selection-font-face-parse.html:
fast/text/font-selection-font-loading-api-parse-expected.txt:
fast/text/font-selection-font-loading-api-parse.html:
fast/text/font-stretch-parse-expected.txt:
fast/text/font-stretch-parse.html:
fast/text/font-style-parse-expected.txt:
fast/text/font-style-parse.html:
fast/text/variations/font-selection-properties.html:
platform/mac-elcapitan/fast/text/font-selection-font-face-parse-expected.txt:

10:48 AM Changeset in webkit [214418] by commit-queue@webkit.org

2 edits in trunk/Source/WebCore

Further optimize checkWebRTCAvailability
https://bugs.webkit.org/show_bug.cgi?id=169147

Patch by Youenn Fablet <youenn@apple.com> on 2017-03-27
Reviewed by Alex Christensen.

Tested locally by removing libwebrtc.dylib.
Replacing dlopen check by checking an exported symbol, rtc::LogMessage::LogToDebug.
This check is more efficient and accurate. It should work in more configurations than the previous one.

platform/mediastream/libwebrtc/LibWebRTCProvider.cpp:

(WebCore::isNullFunctionPointer):
(WebCore::LibWebRTCProvider::webRTCAvailable):

10:21 AM Changeset in webkit [214417] by Jonathan Bedard

2 edits in trunk/Tools

REGRESSION: Failing to iterate over Simulator.managed_devices map
<rdar://problem/31274476>

Unreviewed infrastructure fix.

Scripts/webkitpy/port/ios_simulator.py:

(IOSSimulatorPort._create_devices): Iterate through managed_devices by worker number.

9:39 AM Changeset in webkit [214416] by msaboff@apple.com

3 edits in trunk/Source/JavaScriptCore

Add ARM64 system instructions to disassembler
https://bugs.webkit.org/show_bug.cgi?id=170084

Reviewed by Saam Barati.

This changes adds support for MRS and MSR instructions, and refactors the DMB
disassembly to handle all of the barrier instructions.

disassembler/ARM64/A64DOpcode.cpp:

(JSC::ARM64Disassembler::A64DOpcodeMSRImmediate::format):
(JSC::ARM64Disassembler::A64DOpcodeMSROrMRSRegister::format):
(JSC::ARM64Disassembler::A64DOpcodeSystemSync::format):
(JSC::ARM64Disassembler::A64DOpcodeDmb::format): Deleted.

disassembler/ARM64/A64DOpcode.h:

(JSC::ARM64Disassembler::A64DOpcodeSystem::lBit):
(JSC::ARM64Disassembler::A64DOpcodeSystem::op0):
(JSC::ARM64Disassembler::A64DOpcodeSystem::op1):
(JSC::ARM64Disassembler::A64DOpcodeSystem::crN):
(JSC::ARM64Disassembler::A64DOpcodeSystem::crM):
(JSC::ARM64Disassembler::A64DOpcodeSystem::op2):
(JSC::ARM64Disassembler::A64DOpcodeMSROrMRSRegister::opName):
(JSC::ARM64Disassembler::A64DOpcodeMSROrMRSRegister::systemRegister):
(JSC::ARM64Disassembler::A64DOpcodeSystemSync::opName):
(JSC::ARM64Disassembler::A64DOpcodeSystemSync::option):
(JSC::ARM64Disassembler::A64DOpcodeDmb::opName): Deleted.
(JSC::ARM64Disassembler::A64DOpcodeDmb::option): Deleted.
(JSC::ARM64Disassembler::A64DOpcodeDmb::crM): Deleted.

9:18 AM Changeset in webkit [214415] by mmaxfield@apple.com

3 edits
2 adds in trunk

Implement format specifier for variation fonts
https://bugs.webkit.org/show_bug.cgi?id=169327

Reviewed by Simon Fraser.

Source/WebCore:

Variation fonts require the format("woff-variations"), format("truetype-variations"),
and format("opentype-variations") format specifies in @font-face rules.

Test: fast/text/variations/font-face-format.html

platform/graphics/mac/FontCustomPlatformData.cpp:

(WebCore::FontCustomPlatformData::supportsFormat):

LayoutTests:

fast/text/variations/font-face-format-expected.html: Added.
fast/text/variations/font-face-format.html: Added.

9:12 AM Changeset in webkit [214414] by mmaxfield@apple.com

6 edits
6 adds in trunk

Make sure animation works for font-weight, font-stretch, and font-style
https://bugs.webkit.org/show_bug.cgi?id=169683

Reviewed by Simon Fraser.

Source/WebCore:

Hook up animation code for FontSelectionValues.

Tests: animations/font-variations/font-stretch.html

animations/font-variations/font-style.html
animations/font-variations/font-weight.html

page/animation/CSSPropertyAnimation.cpp:

(WebCore::blendFunc):
(WebCore::CSSPropertyAnimationWrapperMap::CSSPropertyAnimationWrapperMap):

rendering/style/RenderStyle.cpp:

(WebCore::RenderStyle::setFontWeight):
(WebCore::RenderStyle::setFontStretch):
(WebCore::RenderStyle::setFontItalic):

rendering/style/RenderStyle.h:

(WebCore::RenderStyle::fontWeight):
(WebCore::RenderStyle::fontStretch):
(WebCore::RenderStyle::fontItalic):

LayoutTests:

animations/font-variations/font-stretch-expected.txt: Added.
animations/font-variations/font-stretch.html: Added.
animations/font-variations/font-style-expected.txt: Added.
animations/font-variations/font-style.html: Added.
animations/font-variations/font-weight-expected.txt: Added.
animations/font-variations/font-weight.html: Added.
animations/resources/animation-test-helpers.js:

(compareFontStyle):
(getPropertyValue):
(comparePropertyValue):

8:43 AM Changeset in webkit [214413] by achristensen@apple.com

33 edits
9 adds in trunk

Make WebSockets work in network process
https://bugs.webkit.org/show_bug.cgi?id=169930

Reviewed by Youenn Fablet.

Source/WebCore:

Covered by 136 existing tests in http/tests/websocket/tests/hybi
This also does fine with the 544 websocket tests in the web-platform-tests which we have not yet imported.
Also added http/tests/websocket/tests/hybi/network-process-crash-error.html
to test a new condition that couldn't happen before this move: the NetworkProcess crashing.

Modules/websockets/WebSocketChannel.cpp:

(WebCore::WebSocketChannel::fail):
We were asserting that didCloseSocketStream was called. It is still called,
but not synchronously like it used to. This assertion is now invalid, but tests
that would hit it still pass.

platform/network/cf/SocketStreamHandleImplCFNet.cpp:

(WebCore::SocketStreamHandleImpl::platformSendInternal):
CFWriteStreamCanAcceptBytes crashes if you give it a null parameter, and that can happen now.
If we have no write stream, then we cannot write. Tests that hit this pass still.

Source/WebKit2:

This adds WebSocketStream, which is a SocketStreamHandle that sends asynchronous
messages to a NetworkSocketStream in the NetworkProcess. WebKit1 uses a SocketStreamHandleImpl
still, but WebKit2 now has a SocketStreamHandleImpl in the NetworkProcess owned by the
NetworkSocketStream, which is a SocketStreamHandleClient that sends asynchronous messages back
to the WebProcess. WebSocketStream's implementation of SocketStreamHandle::platformSend is special
because it stores the completion handler in a map and asynchronously reports to the caller whether
it succeeded or not. If the network processes crashes before the completion handlers are called,
then they are called indicating failure and an error is reported, but that never happens, right?

CMakeLists.txt:
DerivedSources.make:
NetworkProcess/NetworkConnectionToWebProcess.cpp:

(WebKit::NetworkConnectionToWebProcess::didReceiveMessage):
(WebKit::NetworkConnectionToWebProcess::createSocketStream):
(WebKit::NetworkConnectionToWebProcess::destroySocketStream):

NetworkProcess/NetworkConnectionToWebProcess.h:
NetworkProcess/NetworkConnectionToWebProcess.messages.in:
NetworkProcess/NetworkProcess.cpp:

(WebKit::NetworkProcess::setAllowsAnySSLCertificateForWebSocket):

NetworkProcess/NetworkProcess.h:
NetworkProcess/NetworkProcess.messages.in:
NetworkProcess/NetworkSocketStream.cpp: Added.

(WebKit::NetworkSocketStream::create):
(WebKit::NetworkSocketStream::NetworkSocketStream):
(WebKit::NetworkSocketStream::sendData):
(WebKit::NetworkSocketStream::close):
(WebKit::NetworkSocketStream::didOpenSocketStream):
(WebKit::NetworkSocketStream::didCloseSocketStream):
(WebKit::NetworkSocketStream::didReceiveSocketStreamData):
(WebKit::NetworkSocketStream::didFailToReceiveSocketStreamData):
(WebKit::NetworkSocketStream::didUpdateBufferedAmount):
(WebKit::NetworkSocketStream::didFailSocketStream):
(WebKit::NetworkSocketStream::messageSenderConnection):
(WebKit::NetworkSocketStream::messageSenderDestinationID):

NetworkProcess/NetworkSocketStream.h: Added.
NetworkProcess/NetworkSocketStream.messages.in: Added.
UIProcess/API/C/WKContext.cpp:

(WKContextSetAllowsAnySSLCertificateForWebSocketTesting):
(WKContextTerminateNetworkProcess):

UIProcess/API/C/WKContextPrivate.h:
UIProcess/WebProcessPool.cpp:

(WebKit::WebProcessPool::sendSyncToNetworkingProcess):
(WebKit::WebProcessPool::setAllowsAnySSLCertificateForWebSocket):
(WebKit::WebProcessPool::terminateNetworkProcess):

UIProcess/WebProcessPool.h:
WebKit2.xcodeproj/project.pbxproj:
WebProcess/Network/NetworkProcessConnection.cpp:

(WebKit::NetworkProcessConnection::didReceiveMessage):

WebProcess/Network/WebSocketProvider.cpp:

(WebKit::WebSocketProvider::createSocketStreamHandle):

WebProcess/Network/WebSocketStream.cpp: Added.

(WebKit::globalWebSocketStreamMap):
(WebKit::WebSocketStream::streamWithIdentifier):
(WebKit::WebSocketStream::networkProcessCrashed):
(WebKit::WebSocketStream::create):
(WebKit::WebSocketStream::WebSocketStream):
(WebKit::WebSocketStream::~WebSocketStream):
(WebKit::WebSocketStream::messageSenderConnection):
(WebKit::WebSocketStream::messageSenderDestinationID):
(WebKit::WebSocketStream::platformSend):
(WebKit::WebSocketStream::didSendData):
(WebKit::WebSocketStream::platformClose):
(WebKit::WebSocketStream::bufferedAmount):
(WebKit::WebSocketStream::didOpenSocketStream):
(WebKit::WebSocketStream::didCloseSocketStream):
(WebKit::WebSocketStream::didReceiveSocketStreamData):
(WebKit::WebSocketStream::didFailToReceiveSocketStreamData):
(WebKit::WebSocketStream::didUpdateBufferedAmount):
(WebKit::WebSocketStream::didFailSocketStream):

WebProcess/Network/WebSocketStream.h: Added.
WebProcess/Network/WebSocketStream.messages.in: Added.
WebProcess/WebProcess.cpp:

(WebKit::WebProcess::networkProcessConnectionClosed):

Tools:

WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:

Add terminateNetworkProcess for the new test
http/tests/websocket/tests/hybi/network-process-crash-error.html

WebKitTestRunner/InjectedBundle/TestRunner.cpp:

(WTR::TestRunner::setAllowsAnySSLCertificate):
(WTR::TestRunner::terminateNetworkProcess):

WebKitTestRunner/InjectedBundle/TestRunner.h:
WebKitTestRunner/TestController.cpp:

(WTR::TestController::setAllowsAnySSLCertificate):
(WTR::TestController::terminateNetworkProcess):

WebKitTestRunner/TestController.h:
WebKitTestRunner/TestInvocation.cpp:

(WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
(WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):

WebKitTestRunner/cocoa/TestControllerCocoa.mm:

(WTR::TestController::platformContext):

WebKitTestRunner/gtk/TestControllerGtk.cpp:

(WTR::TestController::platformContext):
(WTR::TestController::platformLibraryPathForTesting):
Add some infrastructure for http/tests/websocket/tests/hybi/simple-wss.html
which calls testRunner.setAllowsAnySSLCertificate. It used to only be effective
for the WebProcess, but SocketStreamHandleImpl::createStreams is now in the NetworkProcess,
and we need its call to Settings::allowsAnySSLCertificate to be loosened for this test,
which tests that wss works, but our test certificate has an invalid certificate chain.
We want production software to not have the ability to have its security weakened, so this
is a test code path that is only implemented in WebKitTestRunner.

LayoutTests:

http/tests/websocket/tests/hybi/network-process-crash-error-expected.txt: Added.
http/tests/websocket/tests/hybi/network-process-crash-error.html: Added.
http/tests/websocket/tests/hybi/slow-reply_wsh.py: Added.

(web_socket_do_extra_handshake):
(web_socket_transfer_data):

platform/ios-simulator-wk1/TestExpectations:
platform/mac-wk1/TestExpectations:
platform/win/TestExpectations:

6:38 AM Changeset in webkit [214412] by commit-queue@webkit.org

3 edits in trunk/Source/WebKit2

[GTK] NetscapePluginX11::m_windowID is unused/unneeded
https://bugs.webkit.org/show_bug.cgi?id=170115

Patch by Adrian Perez de Castro <Adrian Perez de Castro> on 2017-03-27
Reviewed by Michael Catanzaro.

WebProcess/Plugins/Netscape/x11/NetscapePluginX11.cpp:

(WebKit::NetscapePluginX11::NetscapePluginX11):
Remove the unneeded NetscapePluginX11::m_windowID member.

WebProcess/Plugins/Netscape/x11/NetscapePluginX11.h:

Remove the unneeded NetscapePluginX11::m_windowID member.

2:37 AM Changeset in webkit [214411] by commit-queue@webkit.org

11 edits
2 copies
17 adds in trunk

[Modern Media Controls] Improve media documents across macOS, iPhone and iPad
https://bugs.webkit.org/show_bug.cgi?id=169145
<rdar://problem/17048858>

Patch by Antoine Quint <Antoine Quint> on 2017-03-27
Reviewed by Dean Jackson.

Source/WebCore:

There were a variety of issues with media documents, some longstanding, and some specifically
about modern media controls.

One issue was that fullscreen and picture-in-picture buttons would show for audio media documents,
due to using a <video> element to load the audio file. We now have additional logic in MediaController
to identify if the loaded media is really an audio file, and using this information to hide the
fullscreen and picture-in-picture buttons.

Another issue was that we would inject style in MediaDocument.cpp that was specific to modern media
controls when we could have the modern-media-controls module injected CSS handle this styling. We now
use the injected style in the shadow root to size media documents based on the device characteristics
and ensuring that page styles are overridden.

We also simplify how MediaDocument.cpp sets the source of the media element to simply use the "src"
attribute and not a <source> element.

Finally, we introduce a MediaDocumentController class that is instantiated when we're dealing with
a media document to hide the controls while we determine the type of media we're loading (audio vs.
video) in order to apply the appropriate styling without flashes.

As a result of the new styles applied by the modern-media-controls module, media documents have a
similar behavior on macOS and iPad, where we only enforce a min-width for video allowing them
to play at their natural size otherwise, and enforcing a fixed width for audio. On iPhone however,
we want to always play the media at full width, with some padding in the case of audio.

Tests: media/modern-media-controls/fullscreen-support/fullscreen-support-disabled-video-with-audio-tracks-only.html

media/modern-media-controls/media-documents/ipad/media-document-audio-ios-sizing.html
media/modern-media-controls/media-documents/ipad/media-document-video-ios-sizing.html
media/modern-media-controls/media-documents/media-document-audio-ios-sizing.html
media/modern-media-controls/media-documents/media-document-audio-mac-sizing.html
media/modern-media-controls/media-documents/media-document-video-ios-sizing.html
media/modern-media-controls/media-documents/media-document-video-mac-sizing.html
media/modern-media-controls/pip-support/pip-support-disabled-video-with-audio-tracks-only.html

Modules/modern-media-controls/controls/ios-inline-media-controls.css:

(:host(audio) .media-controls.ios.inline > .controls-bar:before,):
(:host(audio) .media-controls.ios.inline > .controls-bar:before): Deleted.

Modules/modern-media-controls/controls/macos-media-controls.css:

(:host(audio) .media-controls.mac.inline > .controls-bar,):
(:host(audio) .media-controls.mac.inline > .controls-bar > .background-tint,):
(:host(audio) .media-controls.mac.inline > .controls-bar): Deleted.
(:host(audio) .media-controls.mac.inline > .controls-bar > .background-tint): Deleted.

Modules/modern-media-controls/controls/media-document.css: Added.

(:host(.media-document)):
(:host(.media-document.ready)):
(:host(.media-document.audio.mac)):
(:host(.media-document.audio.ipad)):
(:host(.media-document.audio.iphone)):
(:host(.media-document.video.mac)):
(:host(.media-document.video.ipad)):
(:host(.media-document.video.iphone)):

Modules/modern-media-controls/js-files:
Modules/modern-media-controls/media/fullscreen-support.js:

(FullscreenSupport.prototype.syncControl):
(FullscreenSupport):

Modules/modern-media-controls/media/media-controller.js:

(MediaController):
(MediaController.prototype.get isAudio):

Modules/modern-media-controls/media/media-document-controller.js: Added.

(MediaDocumentController):
(MediaDocumentController.prototype.handleEvent):
(MediaDocumentController.prototype._mediaDocumentHasMetadata):
(MediaDocumentController.prototype._mediaDocumentHasSize):

Modules/modern-media-controls/media/pip-support.js:

(PiPSupport.prototype.syncControl):
(PiPSupport):

html/MediaDocument.cpp:

(WebCore::MediaDocumentParser::createDocumentStructure):

LayoutTests:

We add new tests for media documents and related features that cover the following cases:

checking <video> with only audio tracks does not show the fullscreen button
checking <video> with only audio tracks does not show the picture-in-picture button
checking the size used in media documents for audio and video across macOS, iPhone and iPad

media/modern-media-controls/fullscreen-support/fullscreen-support-disabled-video-with-audio-tracks-only-expected.txt: Added.
media/modern-media-controls/fullscreen-support/fullscreen-support-disabled-video-with-audio-tracks-only.html: Added.
media/modern-media-controls/media-documents/ipad/media-document-audio-ios-sizing-expected.txt: Added.
media/modern-media-controls/media-documents/ipad/media-document-audio-ios-sizing.html: Added.
media/modern-media-controls/media-documents/ipad/media-document-video-ios-sizing-expected.txt: Added.
media/modern-media-controls/media-documents/ipad/media-document-video-ios-sizing.html: Added.
media/modern-media-controls/media-documents/media-document-audio-ios-sizing-expected.txt: Added.
media/modern-media-controls/media-documents/media-document-audio-ios-sizing.html: Added.
media/modern-media-controls/media-documents/media-document-audio-mac-sizing-expected.txt: Added.
media/modern-media-controls/media-documents/media-document-audio-mac-sizing.html: Added.
media/modern-media-controls/media-documents/media-document-video-ios-sizing-expected.txt: Added.
media/modern-media-controls/media-documents/media-document-video-ios-sizing.html: Added.
media/modern-media-controls/media-documents/media-document-video-mac-sizing-expected.txt: Added.
media/modern-media-controls/media-documents/media-document-video-mac-sizing.html: Added.
media/modern-media-controls/pip-support/pip-support-disabled-video-with-audio-tracks-only-expected.txt: Added.
media/modern-media-controls/pip-support/pip-support-disabled-video-with-audio-tracks-only.html: Added.
platform/ios-simulator/TestExpectations:
platform/mac/TestExpectations:

Mar 26, 2017:

9:17 PM Changeset in webkit [214410] by fpizlo@apple.com

11 edits
4 adds in trunk/Source

B3::fixSSA should do liveness pruning
https://bugs.webkit.org/show_bug.cgi?id=170111

Reviewed by Saam Barati.

Source/JavaScriptCore:

This moves all of the logic of Air::Liveness<> to WTF::Liveness<> and then uses that to
create B3::VariableLiveness. Then this uses VariableLiveness::LiveAtHead to prune Phi
construction.

This makes B3::fixSSA run twice as fast. This is a 13% progression on WasmBench compile
times.

CMakeLists.txt:
JavaScriptCore.xcodeproj/project.pbxproj:
b3/B3BasicBlock.h:

(JSC::B3::BasicBlock::get):

b3/B3FixSSA.cpp:

(JSC::B3::fixSSA):

b3/B3VariableLiveness.cpp: Added.

(JSC::B3::VariableLiveness::VariableLiveness):
(JSC::B3::VariableLiveness::~VariableLiveness):

b3/B3VariableLiveness.h: Added.

(JSC::B3::VariableLivenessAdapter::VariableLivenessAdapter):
(JSC::B3::VariableLivenessAdapter::numIndices):
(JSC::B3::VariableLivenessAdapter::valueToIndex):
(JSC::B3::VariableLivenessAdapter::indexToValue):
(JSC::B3::VariableLivenessAdapter::blockSize):
(JSC::B3::VariableLivenessAdapter::forEachEarlyUse):
(JSC::B3::VariableLivenessAdapter::forEachLateUse):
(JSC::B3::VariableLivenessAdapter::forEachEarlyDef):
(JSC::B3::VariableLivenessAdapter::forEachLateDef):

b3/air/AirCFG.h: Added.

(JSC::B3::Air::CFG::CFG):
(JSC::B3::Air::CFG::root):
(JSC::B3::Air::CFG::newMap):
(JSC::B3::Air::CFG::successors):
(JSC::B3::Air::CFG::predecessors):
(JSC::B3::Air::CFG::index):
(JSC::B3::Air::CFG::node):
(JSC::B3::Air::CFG::numNodes):
(JSC::B3::Air::CFG::dump):

b3/air/AirCode.cpp:

(JSC::B3::Air::Code::Code):

b3/air/AirCode.h:

(JSC::B3::Air::Code::cfg):

b3/air/AirLiveness.h:

(JSC::B3::Air::LivenessAdapter::LivenessAdapter):
(JSC::B3::Air::LivenessAdapter::blockSize):
(JSC::B3::Air::LivenessAdapter::forEachEarlyUse):
(JSC::B3::Air::LivenessAdapter::forEachLateUse):
(JSC::B3::Air::LivenessAdapter::forEachEarlyDef):
(JSC::B3::Air::LivenessAdapter::forEachLateDef):
(JSC::B3::Air::TmpLivenessAdapter::TmpLivenessAdapter):
(JSC::B3::Air::TmpLivenessAdapter::numIndices):
(JSC::B3::Air::StackSlotLivenessAdapter::StackSlotLivenessAdapter):
(JSC::B3::Air::StackSlotLivenessAdapter::numIndices):
(JSC::B3::Air::StackSlotLivenessAdapter::indexToValue):
(JSC::B3::Air::Liveness::Liveness):
(JSC::B3::Air::Liveness::LocalCalc::LocalCalc): Deleted.
(JSC::B3::Air::Liveness::LocalCalc::Iterable::Iterable): Deleted.
(JSC::B3::Air::Liveness::LocalCalc::Iterable::iterator::iterator): Deleted.
(JSC::B3::Air::Liveness::LocalCalc::Iterable::iterator::operator++): Deleted.
(JSC::B3::Air::Liveness::LocalCalc::Iterable::iterator::operator*): Deleted.
(JSC::B3::Air::Liveness::LocalCalc::Iterable::iterator::operator==): Deleted.
(JSC::B3::Air::Liveness::LocalCalc::Iterable::iterator::operator!=): Deleted.
(JSC::B3::Air::Liveness::LocalCalc::Iterable::begin): Deleted.
(JSC::B3::Air::Liveness::LocalCalc::Iterable::end): Deleted.
(JSC::B3::Air::Liveness::LocalCalc::Iterable::contains): Deleted.
(JSC::B3::Air::Liveness::LocalCalc::live): Deleted.
(JSC::B3::Air::Liveness::LocalCalc::isLive): Deleted.
(JSC::B3::Air::Liveness::LocalCalc::execute): Deleted.
(JSC::B3::Air::Liveness::rawLiveAtHead): Deleted.
(JSC::B3::Air::Liveness::Iterable::Iterable): Deleted.
(JSC::B3::Air::Liveness::Iterable::iterator::iterator): Deleted.
(JSC::B3::Air::Liveness::Iterable::iterator::operator*): Deleted.
(JSC::B3::Air::Liveness::Iterable::iterator::operator++): Deleted.
(JSC::B3::Air::Liveness::Iterable::iterator::operator==): Deleted.
(JSC::B3::Air::Liveness::Iterable::iterator::operator!=): Deleted.
(JSC::B3::Air::Liveness::Iterable::begin): Deleted.
(JSC::B3::Air::Liveness::Iterable::end): Deleted.
(JSC::B3::Air::Liveness::Iterable::contains): Deleted.
(JSC::B3::Air::Liveness::liveAtHead): Deleted.
(JSC::B3::Air::Liveness::liveAtTail): Deleted.
(JSC::B3::Air::Liveness::workset): Deleted.

Source/WTF:

Move Air::Liveness<> to WTF::Liveness<>. This is pretty easy since Air::Liveness<> was
already fairly generic. It leverages the CFG concept so that it can understand many
different kinds of basic blocks. It doesn't try to understand the contents of basic
blocks; it just asks the adaptor for the block size and asks for the early/late
uses/defs of each thing in the block.

This makes it easy to create a B3::VariableLiveness, which fixSSA then uses for
pruning. One of the new features is the Liveness::LiveAtHead nested class, which you
instantiate if you want to perform liveAtHead queries, which SSA construction wants to
do.

This also fixes https://bugs.webkit.org/show_bug.cgi?id=170102#c12

WTF.xcodeproj/project.pbxproj:
wtf/CMakeLists.txt:
wtf/Liveness.h: Added.

(WTF::Liveness::Liveness):
(WTF::Liveness::LocalCalc::LocalCalc):
(WTF::Liveness::LocalCalc::Iterable::Iterable):
(WTF::Liveness::LocalCalc::Iterable::iterator::iterator):
(WTF::Liveness::LocalCalc::Iterable::iterator::operator++):
(WTF::Liveness::LocalCalc::Iterable::iterator::operator*):
(WTF::Liveness::LocalCalc::Iterable::iterator::operator==):
(WTF::Liveness::LocalCalc::Iterable::iterator::operator!=):
(WTF::Liveness::LocalCalc::Iterable::begin):
(WTF::Liveness::LocalCalc::Iterable::end):
(WTF::Liveness::LocalCalc::Iterable::contains):
(WTF::Liveness::LocalCalc::live):
(WTF::Liveness::LocalCalc::isLive):
(WTF::Liveness::LocalCalc::execute):
(WTF::Liveness::rawLiveAtHead):
(WTF::Liveness::Iterable::Iterable):
(WTF::Liveness::Iterable::iterator::iterator):
(WTF::Liveness::Iterable::iterator::operator*):
(WTF::Liveness::Iterable::iterator::operator++):
(WTF::Liveness::Iterable::iterator::operator==):
(WTF::Liveness::Iterable::iterator::operator!=):
(WTF::Liveness::Iterable::begin):
(WTF::Liveness::Iterable::end):
(WTF::Liveness::Iterable::contains):
(WTF::Liveness::liveAtHead):
(WTF::Liveness::liveAtTail):
(WTF::Liveness::workset):
(WTF::Liveness::LiveAtHead::LiveAtHead):
(WTF::Liveness::LiveAtHead::isLiveAtHead):

3:11 PM Changeset in webkit [214409] by fpizlo@apple.com

11 edits in trunk/Source

Air::Liveness shouldn't need HashSets
https://bugs.webkit.org/show_bug.cgi?id=170102

Reviewed by Yusuke Suzuki.
Source/JavaScriptCore:

This converts Air::Liveness<> to no longer use HashSets or BitVectors. This turns out to be
easy because it's cheap enough to do a sorted merge of the things being added to liveAtHead and
the things in the predecessors' liveAtTail. This turns out to be faster - it's a 2% overall
compile time progression on WasmBench.

b3/B3LowerToAir.cpp:

(JSC::B3::Air::LowerToAir::lower): Add a FIXME unrelated to this patch.

b3/air/AirLiveness.h:

(JSC::B3::Air::AbstractLiveness::AbstractLiveness):
(JSC::B3::Air::AbstractLiveness::LocalCalc::LocalCalc):
(JSC::B3::Air::AbstractLiveness::rawLiveAtHead):
(JSC::B3::Air::AbstractLiveness::liveAtHead):
(JSC::B3::Air::AbstractLiveness::liveAtTail):

b3/air/AirTmp.h:

(JSC::B3::Air::Tmp::bank):
(JSC::B3::Air::Tmp::tmpIndex):

dfg/DFGStoreBarrierClusteringPhase.cpp:

Source/WTF:

wtf/IndexSparseSet.h: Add some helpers for a HashSet-free liveness analysis.

(WTF::IndexSparseSet::values):
(WTF::IndexSparseSet<OverflowHandler>::sort):

wtf/StdLibExtras.h:

(WTF::mergeDeduplicatedSorted): Rapidly merge two sorted lists that don't have duplicates to produce a new sorted list that doesn't have duplicates.

wtf/Vector.h:

(WTF::minCapacity>::uncheckedAppend): Inline this!
(WTF::removeRepeatedElements): This is a version of std::unique() that works naturally for Vectors.

2:39 PM Changeset in webkit [214408] by fpizlo@apple.com

14 edits
2 adds in trunk/Source

Air should use RegisterSet for RegLiveness
https://bugs.webkit.org/show_bug.cgi?id=170108

Reviewed by Yusuke Suzuki.
Source/JavaScriptCore:

The biggest change here is the introduction of the new RegLiveness class. This is a
drop-in replacement for the old RegLiveness, which was a specialization of
AbstractLiveness<>, but it's about 30% faster. It gets its speed boost from just using
sets everywhere, which is efficient for registers since RegisterSet is just two (on
x86-64) or three 32-bit (on ARM64) statically allocated words. This looks like a 1%
compile time progression on WasmBench.

CMakeLists.txt:
JavaScriptCore.xcodeproj/project.pbxproj:
b3/B3TimingScope.cpp: Records phase timing totals.

(JSC::B3::TimingScope::TimingScope):
(JSC::B3::TimingScope::~TimingScope):

b3/B3TimingScope.h:
b3/air/AirAllocateRegistersByGraphColoring.cpp:

(JSC::B3::Air::allocateRegistersByGraphColoring):

b3/air/AirLiveness.h: Move code around and rename a bit to make it more like RegLiveness; in particular we want the iterator to be called iterator not Iterator, and we want it to be internal to its iterable. Also rename this template to Liveness, to match the header filename.

(JSC::B3::Air::Liveness::Liveness):
(JSC::B3::Air::Liveness::LocalCalc::LocalCalc):
(JSC::B3::Air::Liveness::LocalCalc::Iterable::Iterable):
(JSC::B3::Air::Liveness::LocalCalc::Iterable::iterator::iterator):
(JSC::B3::Air::Liveness::LocalCalc::Iterable::iterator::operator++):
(JSC::B3::Air::Liveness::LocalCalc::Iterable::iterator::operator*):
(JSC::B3::Air::Liveness::LocalCalc::Iterable::iterator::operator==):
(JSC::B3::Air::Liveness::LocalCalc::Iterable::iterator::operator!=):
(JSC::B3::Air::Liveness::LocalCalc::Iterable::begin):
(JSC::B3::Air::Liveness::LocalCalc::Iterable::end):
(JSC::B3::Air::Liveness::Iterable::Iterable):
(JSC::B3::Air::Liveness::Iterable::iterator::iterator):
(JSC::B3::Air::RegLivenessAdapter::RegLivenessAdapter): Deleted.
(JSC::B3::Air::RegLivenessAdapter::numIndices): Deleted.
(JSC::B3::Air::RegLivenessAdapter::acceptsBank): Deleted.
(JSC::B3::Air::RegLivenessAdapter::acceptsRole): Deleted.
(JSC::B3::Air::RegLivenessAdapter::valueToIndex): Deleted.
(JSC::B3::Air::RegLivenessAdapter::indexToValue): Deleted.
(JSC::B3::Air::AbstractLiveness::AbstractLiveness): Deleted.
(JSC::B3::Air::AbstractLiveness::LocalCalc::LocalCalc): Deleted.
(JSC::B3::Air::AbstractLiveness::LocalCalc::Iterator::Iterator): Deleted.
(JSC::B3::Air::AbstractLiveness::LocalCalc::Iterator::operator++): Deleted.
(JSC::B3::Air::AbstractLiveness::LocalCalc::Iterator::operator*): Deleted.
(JSC::B3::Air::AbstractLiveness::LocalCalc::Iterator::operator==): Deleted.
(JSC::B3::Air::AbstractLiveness::LocalCalc::Iterator::operator!=): Deleted.
(JSC::B3::Air::AbstractLiveness::LocalCalc::Iterable::Iterable): Deleted.
(JSC::B3::Air::AbstractLiveness::LocalCalc::Iterable::begin): Deleted.
(JSC::B3::Air::AbstractLiveness::LocalCalc::Iterable::end): Deleted.
(JSC::B3::Air::AbstractLiveness::LocalCalc::Iterable::contains): Deleted.
(JSC::B3::Air::AbstractLiveness::LocalCalc::live): Deleted.
(JSC::B3::Air::AbstractLiveness::LocalCalc::isLive): Deleted.
(JSC::B3::Air::AbstractLiveness::LocalCalc::execute): Deleted.
(JSC::B3::Air::AbstractLiveness::rawLiveAtHead): Deleted.
(JSC::B3::Air::AbstractLiveness::Iterable::Iterable): Deleted.
(JSC::B3::Air::AbstractLiveness::Iterable::iterator::iterator): Deleted.
(JSC::B3::Air::AbstractLiveness::Iterable::iterator::operator*): Deleted.
(JSC::B3::Air::AbstractLiveness::Iterable::iterator::operator++): Deleted.
(JSC::B3::Air::AbstractLiveness::Iterable::iterator::operator==): Deleted.
(JSC::B3::Air::AbstractLiveness::Iterable::iterator::operator!=): Deleted.
(JSC::B3::Air::AbstractLiveness::Iterable::begin): Deleted.
(JSC::B3::Air::AbstractLiveness::Iterable::end): Deleted.
(JSC::B3::Air::AbstractLiveness::Iterable::contains): Deleted.
(JSC::B3::Air::AbstractLiveness::liveAtHead): Deleted.
(JSC::B3::Air::AbstractLiveness::liveAtTail): Deleted.
(JSC::B3::Air::AbstractLiveness::workset): Deleted.

b3/air/AirLogRegisterPressure.cpp:
b3/air/AirLowerAfterRegAlloc.cpp:
b3/air/AirRegLiveness.cpp: Added.

(JSC::B3::Air::RegLiveness::RegLiveness):
(JSC::B3::Air::RegLiveness::~RegLiveness):
(JSC::B3::Air::RegLiveness::LocalCalc::execute):

b3/air/AirRegLiveness.h: Added.

(JSC::B3::Air::RegLiveness::LocalCalc::LocalCalc):
(JSC::B3::Air::RegLiveness::LocalCalc::live):
(JSC::B3::Air::RegLiveness::LocalCalc::isLive):
(JSC::B3::Air::RegLiveness::liveAtHead):
(JSC::B3::Air::RegLiveness::liveAtTail):

b3/air/AirReportUsedRegisters.cpp:
jit/RegisterSet.h:

(JSC::RegisterSet::add):
(JSC::RegisterSet::remove):
(JSC::RegisterSet::contains):
(JSC::RegisterSet::subsumes):
(JSC::RegisterSet::iterator::iterator):
(JSC::RegisterSet::iterator::operator*):
(JSC::RegisterSet::iterator::operator++):
(JSC::RegisterSet::iterator::operator==):
(JSC::RegisterSet::iterator::operator!=):
(JSC::RegisterSet::begin):
(JSC::RegisterSet::end):

Source/WTF:

wtf/Atomics.h:

(WTF::ensurePointer): This is a useful replacement for std::once, which requires less fencing.

wtf/Bitmap.h: Add more utilities.

(WTF::Bitmap::iterator::iterator): An iterator for set bits.
(WTF::Bitmap::iterator::operator*):
(WTF::Bitmap::iterator::operator++):
(WTF::Bitmap::iterator::operator==):
(WTF::Bitmap::iterator::operator!=):
(WTF::Bitmap::begin):
(WTF::Bitmap::end):
(WTF::WordType>::subsumes): a.subsumes(b) if all of b's set bits are set in a.
(WTF::WordType>::findBit): find next set or clear bit.

Note: See TracTimeline for information about the timeline view.

Download in other formats:

RSS Feed