[webkit-changes] [WebKit/WebKit] b99d74: Cherry-pick 3a3a97e3b214. rdar://problem/100814764

Antoine Quint noreply at github.com
Thu Oct 6 13:15:19 PDT 2022


  Branch: refs/heads/safari-7615.1.8-branch
  Home:   https://github.com/WebKit/WebKit
  Commit: b99d74a41f06c44b9204d4273a2776a266088acd
      https://github.com/WebKit/WebKit/commit/b99d74a41f06c44b9204d4273a2776a266088acd
  Author: Antoine Quint <graouts at webkit.org>
  Date:   2022-10-06 (Thu, 06 Oct 2022)

  Changed paths:
    M LayoutTests/imported/w3c/web-platform-tests/css/css-pseudo/first-letter-allowed-properties-expected.txt
    M LayoutTests/imported/w3c/web-platform-tests/css/css-pseudo/first-line-allowed-properties-expected.txt
    M LayoutTests/imported/w3c/web-platform-tests/css/css-transforms/transform-2d-getComputedStyle-001.html
    M Source/WebCore/animation/KeyframeEffect.cpp
    M Source/WebCore/animation/WebAnimation.cpp
    M Source/WebCore/css/ComputedStyleExtractor.cpp
    M Source/WebCore/css/ComputedStyleExtractor.h

  Log Message:
  -----------
  Cherry-pick 3a3a97e3b214. rdar://problem/100814764

    REGRESSION(254760 at main): mobile version of apple.com/iphone-14 doesn't display
    https://bugs.webkit.org/show_bug.cgi?id=246122
    rdar://100814764

    Reviewed by Antti Koivisto.

    We used to identify a null renderer in computedTransform() as a signal that the "computed" value
    for "transform" was requested rather than the "resolved" value. The difference is significant
    because a the "resolved" value for "transform" for an element that has "display: none", and thus
    no renderer is "none". However, we failed to identify this case and would return the <transform-list>
    type.

    The same way that ComputedStyleExtractor::propertyValue() takes in an argument to specify the value
    type requested, we add a new argument to ComputedStyleExtractor::valueForPropertyInStyle() which we
    pass to computedTransform() to distinguish between the "display: none" case and the "computed" value
    case.

    We also ask for the "computed" value from WebAnimation::commitStyles() and when parsing keyframe styles
    in the Web Animations API. We were already requesting such style through ComputedStyleExtractor::propertyValue()
    in the CSS Typed OM computedStyleMap() function.

    Since this case was not tested before, we add some new assertions in an existing WPT test to check that
    setting "display: none" on an element with a "transform" set does indeed return "none".

    Incidentally, this brought a couple of new WPT PASS results for ::first-letter and ::first-line.

    * LayoutTests/imported/w3c/web-platform-tests/css/css-pseudo/first-letter-allowed-properties-expected.txt:
    * LayoutTests/imported/w3c/web-platform-tests/css/css-pseudo/first-line-allowed-properties-expected.txt:
    * LayoutTests/imported/w3c/web-platform-tests/css/css-transforms/transform-2d-getComputedStyle-001.html:
    * Source/WebCore/animation/KeyframeEffect.cpp:
    (WebCore::KeyframeEffect::getKeyframes):
    * Source/WebCore/animation/WebAnimation.cpp:
    (WebCore::WebAnimation::commitStyles):
    * Source/WebCore/css/ComputedStyleExtractor.cpp:
    (WebCore::computedTransform):
    (WebCore::ComputedStyleExtractor::propertyValue):
    (WebCore::ComputedStyleExtractor::valueForPropertyInStyle):
    * Source/WebCore/css/ComputedStyleExtractor.h:

    Canonical link: https://commits.webkit.org/255216@main

Canonical link: https://commits.webkit.org/255076.3@safari-7615.1.8-branch




More information about the webkit-changes mailing list