Robert's comment 3 seems like the right diagnosis. See my just-attached further-reduce testcase, where Blink and WebKit applies the same "orange" rule to the source element as well as its clone in the `use` subtree, even though that rule is targeted based on an ancestor-selector for an ancestor that's outside of the use-cloned subtree. Our behavior here (rendering the second rect as `lime`) is correct, per the spec text that Robert quoted.
Bug 1760954 Comment 5 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
Robert's comment 3 seems like the right diagnosis. See my just-attached further-reduce testcase, where Blink and WebKit applies the same "orange" rule to the source element as well as its clone in the `use` subtree, even though that rule is targeted based on an ancestor-selector for an ancestor that's outside of the use-cloned subtree. Our behavior here (rendering the reduced testcase's second rect as `lime`) is correct, per the spec text that Robert quoted.