240563 2022-05-18 00:18:50 -0700 Reflection for FrozenArray<Element> caching invariant 2022-06-02 14:45:56 -0700 1 1 1 Unclassified WebKit DOM WebKit Nightly Build Unspecified Unspecified RESOLVED FIXED https://github.com/web-platform-tests/wpt/pull/34194 InRadar P2 Normal --- 196843 1 rego webkit-unassigned cdumez rniwa webkit-bug-importer oldest_to_newest 1869932 0 rego 2022-05-18 00:18:50 -0700 In bug #239853 we implemented reflection for FrozenArray<Element> attributes following the spec text at: https://github.com/whatwg/html/pull/3917 However the implementation doesn't match one of the things in the spec: > This extra caching layer is necessary to preserve the invariant that element.reflectedElements === element.reflectedElements. We don't have an extra caching layer so we can return the very same array in this case. The contents of the arrays are the same but the previous condition is false. I'm not really sure how we could implemented that, as we're storing the "explicitly set attr-elements" using "HashMap<QualifiedName, Vector<WeakPtr<Element>>>" but Element::getElementsArrayAttribute() returns "Vector<RefPtr<Element>>". I'm not sure if we can cache the latter somehow without causing memory leaks. Domenic suggested that ObservableArray could fix this issue (see https://github.com/web-platform-tests/wpt/pull/34095#discussion_r875333836), but we don't have support yet (bug #238281). Any advice about how we can solve this issue? Thanks! 1870503 1 rego 2022-05-19 12:40:35 -0700 cdumez@ suggested to use JSValueInWrappedObject to fix this issue. I'll give that a try. Thanks. 1871664 2 rego 2022-05-24 22:52:38 -0700 Pull request with the patch: https://github.com/WebKit/WebKit/pull/974 1871675 3 webkit-bug-importer 2022-05-25 00:19:13 -0700 <rdar://problem/93884533> 1873699 4 ews-feeder 2022-06-02 14:45:53 -0700 Committed r295148 (251237@main): <https://commits.webkit.org/251237@main> Reviewed commits have been landed. Closing PR #1166 and removing active labels.