[css-view-transitions-2] Clarify inbound transition timeout semantics

[mattwoodrow]

The current spec mentions a UA defined timeout as part of resolve inbound cross-document view-transition, which then cancels the active view transition (in a queued task).

The pageswap-long-delay.html test is using a very-long synchronous pageswap event to trigger the timeout, which suggests that the timed period starts before this event (which is fairly removed and before the resolve inbound cross-document view-transition algorithm).

It would be good to clarify when the timed period starts.

The test is also expecting the pagereveal event's viewTransition parameter to be null, which is different from queued cancellation of the view transition. I'm not sure if this is a test bug, or a spec bug.

[khushalsagar]

Looks like a spec bug. The spec text is for #9155 which asks to abort the transition if the navigation is too long. The navigation starts while the old Document is still active and pageswap is part of the navigation phase.

Was checking in case we got it wrong from a refactor but looks like this was set up that way since the PR which added it: https://github.com/w3c/csswg-drafts/pull/9608/files.

@noamr can you take a look?

[noamr]

See #10817