Re: [RequestAnimationFrame] Processing model does not permit handling animation callbacks for multiple documents at once

I've pushed a new version of the processing model viewable
https://dvcs.w3.org/hg/webperf/raw-file/7036e94e748c/specs/RequestAnimationFrame/Overview.htmlthat
I believe resolves this issue.  Tasks are now generated and processed
on a top-level browsing contexts rather than on Documents.  Each task first
assembles a list of callbacks for each active document within the top-level
browsing context and then runs them.  The order in which callbacks from
different documents run is not specified, but callbacks on the same
document must stay in the same order.

I do not think this is a significant change for implementations since I
think it's closer to the intended behavior.  A UA may run tasks for
multiple top-level browsing contexts back-to-back without updating the
rendering in between, or allow the rendering to be updated in between tasks
for different top-level browsing contexts.  A conforming UA may not update
the rendering after running callbacks for some but not all Documents within
the same top-level browsing context, but I'm pretty sure nobody would ever
want to actually do that.

- James