W3C home > Mailing lists > Public > public-web-perf@w3.org > February 2013

Re: requestAnimationFrame behavior on display:none iframes

From: James Robinson <jamesr@google.com>
Date: Thu, 14 Feb 2013 16:21:05 -0800
Message-ID: <CAD73mdJbX+Nk_Wt9aM-YyjDfsOHQdeQ9fNry-wK_ib3P9Mx2GQ@mail.gmail.com>
To: Philippe Le Hegaret <plh@w3.org>
Cc: Nat Duca <nduca@chromium.org>, public-web-perf <public-web-perf@w3.org>, Cameron McCormack <cam@mcc.id.au>
On Thu, Feb 14, 2013 at 5:32 AM, Philippe Le Hegaret <plh@w3.org> wrote:

> On Wed, 2013-02-13 at 16:52 -0800, James Robinson wrote:
> >         > Should requestAnimationFrame tick on display:none iframes?
> >         >
> >         > A simple test:
> >         >
> >
> http://www.jankfree.com/raf/iframes_and_visibility_when_hidden.html
> >         >
> >         > FF and Chrome differ:
> >         > -  FF doesn't tick
> >         > - Webkit does
> >         >
> >         > Not sure what IE does.
> >         >
> >         > And, what do we think is "correct"?
> >
> >         Just to make sure we're closing the loop here. Section 5 of
> >         the
> >         processing model [1] indicates that if the context is not
> >         visible
> >         (hidden is true), no callback happens.
> >         See also the response from James in
> >
> http://lists.w3.org/Archives/Public/public-web-perf/2012Oct/0033.html
> >
> > No, that's not correct at all.
> >
> Then, my issues list is incorrect and we haven't really addressed Nat
> Duca's issue.
>
> Can you look at the short thread at
>  http://lists.w3.org/Archives/Public/public-web-perf/2012Jul/0001.html
>
> and see if we need to do anything?
>

He raises an interesting use case that isn't addressed by the current RAF
spec.  To generalize the concern a bit, the issue he's running into is the
browser does not have enough information about what content a web developer
is trying to animate or otherwise manipulate, and authors do not have a
good way to determine what content on a page is or isn't visible (or likely
to soon become visible).  document.hidden, for instance, just punts up to
the top level browsing context for visiblity information (which leads to
the RAF behavior).  We should definitely try to come up with solutions for
this problem either in RAF or in other specifications.  The element
parameter was an attempt to tackle this within RAF itself, but it ran into
difficulties.  It may be that if we provide authors better information
about visibility for iframes and for other forms of DOM content we can come
up with a more general solution.

However, the behavior in the current RAF spec is well-defined and I don't
think we want to attempt to change it for this level.  We should add tests,
make sure current implementations are in compliance, and continue the
general discussion where appropriate.

- James


> Thank you,
>
> Philippe
>
>
>
>
Received on Friday, 15 February 2013 00:21:34 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 15 February 2013 00:21:34 GMT