[cssom-view] Allow elementFromPoint and elementsFromPoint to traverse shadow roots

[captainbrosset]

I recently asked on Mastodon if folks knew the element(s)FromPoint() methods and whether they'd want them to be able to traverse shadow roots: https://mas.to/@patrickbrosset/113236606097090887

Some replies to my "Do you wish it would across Shadow DOM boundaries "question:

In my use case, whether I want it to cross Shadow DOM boundaries would depend on how "thick" the custom elements are. If thin, I wouldn't care. If they encompassed too much in the page, then yeah I'd want a more fine-grained approach.

YES

ABSOLUTELY YES

I'm not sure whether it should cross shadow boundaries on its own. I could maybe see an option, so you had to do work to get through, but didn't need to write your own recursive method.

I think I’d want to be explicit about getting any element from within a shadow root, so maybe if there was an option to include shadow roots.

Absolutely yes! Eventually, it could be a parameter option, like document.elementsFromPoint({ composed: true }) or similar.

Note that caretPositionFromPoint already works this way: https://developer.mozilla.org/en-US/docs/Web/API/Document/caretPositionFromPoint#shadowroots

Also, during the TAG review for caretPositionFromPoint @LeaVerou said (w3ctag/design-reviews#949 (comment)):

I would reiterate that elementFromPoint() could also be expanded in the same way for consistency (and I suspect there are plenty of use cases for that too).

[kbabbitt]

To clarify: Is the proposal to extend elementFromPoint / elementsFromPoint with an optional options parameter which can contain a list of shadowRoots, similar to what we added to caretPositionFromPoint?

[captainbrosset]

To clarify: Is the proposal to extend elementFromPoint / elementsFromPoint with an optional options parameter which can contain a list of shadowRoots, similar to what we added to caretPositionFromPoint?

Correct

[css-meeting-bot]

The CSS Working Group just discussed [cssom-view] Allow elementFromPoint and elementsFromPoint to traverse shadow roots, and agreed to the following:

• RESOLVED: add shadowRoots param to ePFP() to match caretPFP()

The full IRC log of that discussion

<TabAtkins> kbabbitt: request is to have elementFromPoint() be able to traverse shadows, so you can get the "real" element when it's in a shadow
<TabAtkins> astearns: simialr to caretPositionFromPoint(), i see
<emilio> q+
<TabAtkins> proposed: add shadowRoots param to ePFP() to match caretPFP()
<ydaniv> +1
<astearns> ack emilio
<TabAtkins> emilio: somewhat related, Gecko has a way of getting raw nodes from point
<TabAtkins> emilio: text nodes, etc
<TabAtkins> emilio: it always felt kinda funny that elementFromPoint() wouldn't giv eyou that.
<TabAtkins> emilio: wonder if that owuld be useful
<TabAtkins> emilio: nodesFromPoint(), returning the actual node when the point is text, rather than closest element
<TabAtkins> astearns: seems useful to me, open an issue?
<TabAtkins> emilio: we also have the ability to ask what elements/nodes intersect a rect, needed for finger tap testing
<TabAtkins> emilio: I'll file an issue
<TabAtkins> RESOLVED: add shadowRoots param to ePFP() to match caretPFP()