Correct unload event timing for different COOP

[noamr]

Sometimes a navigation causes a BCG swap, e.g. when the
cross-origin isolated capability changes.

In that case:

• the timestamps should coarsened based on both documents
• Some user-agents unload in parallel, which should not become
  part of the new document's timing info
• Time origin should be according to the new global's time origin, but coarsened by both.

Closes w3c/navigation-timing#169

• At least two implementers are interested (and none opposed):
  • …
  • …
• Tests are written and can be reviewed and commented upon at:
  • …
• Implementation bugs are filed:
  • Chrome: …
  • Firefox: …
  • Safari: …
  • Deno (only for timers, structured clone, base64 utils, channel messaging, module resolution, web workers, and web storage): …
  • Node.js (only for timers, structured clone, base64 utils, channel messaging, and module resolution): …

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

---

/browsing-the-web.html ( diff )

[domenic]

It's not clear whether this solution has implementer buy-in, but it looks about right as a spec strategy if it does. I'll comment over there about some implications.

However if we do go this route it'd be helpful if you were also able to make an initial PR against #6315 (i.e. against the session-history branch). I've been merging in changes so far, but I'm not as confident this one can be carried over easily...

[noamr]

However if we do go this route it'd be helpful if you were also able to make an initial PR against #6315 (i.e. against the session-history branch). I've been merging in changes so far, but I'm not as confident this one can be carried over easily...

#7953

[domenic]

Thanks so much for doing #7953 as well. That has similar issues to my review here but you don't need to worry about them. Now that I understand how to integrate, I think I can take over the merge process for that one.