Remove MessagePort's close event

[nicolo-ribaudo]

As discussed in #12957. Note that I have no interest in working on re-adding the event, but the commit description should have enough pointer for somebody else that might.

Closes #10201, closes #12797, closes #12957

Commit message:

Remove MessagePort's close event

This is a revert of cc2634f (it's not a "clean" revert because of some other changes that happened in the meantime).

This means that this patch also restores the older GC semantics: a port is not easily GC-able, it is not exlicitly disentangled when the other half's owner document is destroyed, and relies on polling to see if the other half is still alive. I believe this matches Firefox's behavior, which is also what I believe the discussion in #10201 leads to.

Note that browsers have all different GC semantics (according to Claude):

• Chromium makes a port uncollectable only if it's already started
• Firefox does not collect them
• WebKit keeps the port uncollectable only if it has a message listener

Given that no browser ships the close events this revert is probably the best thing to do; if somebody in the future wants to reintroduce it some relevant discussions are:

• Add onclose event to MessagePort #1766
• MessagePort API:add a close event #9933
• Restrict when MessagePort's onclose event can fire #10201
• MessagePort should probably fire close event from a task #12797
• Fire the "close" event on MessagePort from a task #12957

---

• At least two implementers are interested (and none opposed):
  • N/A, feature did not ship and has known issues.
• Tests are written and can be reviewed and commented upon at:
  • Remove MessagePort close event tests web-platform-tests/wpt#63174
• Implementation bugs are filed:
  • Chromium: https://issues.chromium.org/issues/568711380
  • Gecko: Commented at https://bugzilla.mozilla.org/show_bug.cgi?id=1874820
  • WebKit: Commented at https://bugs.webkit.org/show_bug.cgi?id=267567
  • Deno (only for timers, structured clone, base64 utils, channel messaging, module resolution, web workers, and web storage): Commented at Implementation of close event of MessagePort API denoland/deno#21964
  • Node.js (only for timers, structured clone, base64 utils, channel messaging, and module resolution): MessagePort close event removed from the HTML spec nodejs/node#66460
• Corresponding HTML AAM & ARIA in HTML issues & PRs: Not relevant
• MDN issue is filed: Not documented (Add docs for close event of MessagePort API mdn/mdn#517)
• The top of this comment includes a clear commit message to use.

(See WHATWG Working Mode: Changes for more details.)

---

/document-lifecycle.html ( diff )
/index.html ( diff )
/indices.html ( diff )
/web-messaging.html ( diff )

[nicolo-ribaudo]

The disentangle steps were introduced in the original onclose PR, but it's nice to have a dfn for the various places that disentangle ports.

[zcorpan]

If this matches current implementations, that satisfies the implementer interest item. Should have tests to verify the lack of an event and event handler IDL attribute.

[nicolo-ribaudo]

If this matches current implementations, that satisfies the implementer interest item.

Well, it kind of matches implementations. No implementation ships close (Chrome did implement it, but it's behind I believe a test-only flag).

However all implementation disagree on GC semantics, which this PR restores to the pre-close-event state. None implemented the post-close-event GC behavior, and only Firefox has the old one.

Should have tests to verify the lack of an event and event handler IDL attribute.

Done