[webkit-dev] Expected behavior of "scrolling" attribute for IFrame element

[Ojan Vafai]

On Mon, Jun 13, 2011 at 11:21 AM, Darin Adler <darin at apple.com> wrote:

> On Jun 13, 2011, at 11:05 AM, Ojan Vafai wrote:
> > On Mon, Jun 13, 2011 at 10:51 AM, Darin Adler <darin at apple.com> wrote:
> >> Even though we need to prevent find or autoscrolling from scrolling, it
> seems we must not prevent explicit programmatic scrolling. I’ll add a
> comment to the bug about this.
> >
> > overflow:hidden divs are used to implement custom scrollbars. While
> find-in-page breaking wouldn't be too bad, breaking autoscrolling on these
> sites would likely require us to rollback the change. Maybe we should expose
> an attribute or CSS property to allow controlling this to give sites a
> workaround?
>
> Could be.
>
> I guess at root this is a user interface problem. Scrolling something if
> there is no UI for scrolling back is a problem. Similarly, scrolling
> something into view that a developer thinks is invisible is a problem. I
> guess the browser does need some way to tell if the user can scroll back, so
> it can various forms of scrolling in those cases.
>
> It would be best if we could correctly do this even on existing sites, but
> I don’t have any ideas about how to tell cases where a scroll would be a bad
> idea from cases where a scroll would be OK.
>

I agree. I think it would be fine to disable autoscroll for overflow:hidden
by default if we give a way for sites to turn it back on.

Implementing custom scrollbars is often a bad idea for websites that want to
> work well on platforms like Mac OS X Lion with a modern input device and iOS
> that don’t have conventional scrollbars. Not sure how to promulgate the best
> practices here.
>

Yeah. I can't think of a way to improve the situation beyond just giving
developers control over autoscroll.

Just curious: On the sites you know of that do this to implement custom
> scrollbars, do they use the scroll event to update the scrollbars when
> scrolling is done by the browser engine?
>

I'm not sure. The only real-world case of this I can think of is Google
Wave. I'm not sure what their implementation is like though.


>    -- Darin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-dev/attachments/20110615/f1e2da36/attachment.html>