attr-associated elements use inconsistent trees?

[emilio]

https://html.spec.whatwg.org/multipage/common-dom-interfaces.html#attr-associated-element says:

If reflectedTarget's explicitly set attr-element is a descendant of any of element's shadow-including ancestors, then return reflectedTarget's explicitly set attr-element.

I wonder if the use of DOM descendant but then shadow-including ancestors is intentional? Seems a bit weird that it's not symmetric but maybe it's fine?

[emilio]

cc @mfreed7

[annevk]

That's intentional. As otherwise you break encapsulation.

[emilio]

Might be worth a note in the spec then :)

[emilio]

(Though if you had already associated the element explicitly it seems it would've been fine to just check that both elements are shadow-including-connected? I don't have a strong opinion, it was just a bit weird when reading it)

[annevk]

If you want a note we should keep this open.

The reason is that the getter in question is public API and can thus be observed by scripts otherwise encapsulated from the shadow tree.

I created #8932 to give more power to getters where that is not the case (those on ElementInternals).