as I'm not completely au-fait with what composed events actually are/do, nor if there are any issues with a W3C spec referencing WHATWG (political or otherwise) ... @RByers, thoughts?

The clearest definition of composed seems to be in the note here:

True if event invokes listeners past a ShadowRoot node that is the root of its target attribute value, and false otherwise.

I think the general approach (of extending this spec and referring to the WHATWG DOM spec) is fine. When / if we ever go to publish a CR spec we may need to replace the reference to a W3C version, but presumably by then the W3C fork will be updated to also define composed.

I have a couple questions / concerns with the overall approach though - I'll follow up on WICG/webcomponents#513 and circle back here.

Thank you for the review.

When / if we ever go to publish a CR spec we may need to replace the reference to a W3C version, but presumably by then the W3C fork will be updated to also define composed.

I understand the concern. Actually, I thought the similar thing in creating the PR because it looks that most (all?) of W3C specification do not have a reference to a WHATWG Standard. It looks they try to avoid to have a reference to WHATWG. I am not 100% sure how we proceed in this case.

@annevk do you have any advice? Should I wait W3C version for incorporating Event.composed from WHATWG DOM?

If having a reference to WHATWG DOM is okay, I will try to make a PR for UIEvents at first here w3c/uievents#89. After that, let me ask a merge of PR for Touch Events and Pointer Events, again.

There shouldn't be any issue referencing WHATWG documents. Happens all the time. I don't think anyone is maintaining the DOM fork either (which is annoying, since you have to know that is the case, it would be much easier if they just stopped copy-and-pasting without due process).

@annevk Thank you!

@RByers

PR for UI events has been merged w3c/uievents#91.
Could you review this PR now?

LGTM, thanks!