Gisle Aas wrote:
> We might try to deal with:
>
> Refresh: 0; ...
>
> as if it was a normal 3xx-redirect. If the number is something else
> than 0 then the page should simply be returned as now.
Most Netscape browsers get "stuck" on b.html if a.html does a Refresh with a
zero timeout. When you go BACK from b.html to a.html it slams you forward again.
The user's perception is that the BACK button is broken.
Hence I advise my clients to use a short timeout of 1 or 2 seconds so the user
has at least the chance of seeing what is going on and hitting BACK again to
escape out of a.html. Yes, it's a kludge, but there's still a lot of Netscape
users out there...
Anyway, the proposed patch wouldn't help for a page with a timeout of 1.
What about this:
refresh_ok(0) disable refresh following
refresh_ok(1) follow refreshes of 1 second or less
refresh_ok(60) follow refreshes of 60 seconds or less
And if refresh_ok has not been set (is undef) it could default to the current
value of redirect_ok.
Just a random idea from the peanut gallery...
(Yes I recognize there is no way to say "follow refreshes of 0 seconds only" in
this solution. I thought the 0 would be better as an off switch a la
redirect_ok. I suppose the semantics COULD be "refresh_ok(60) follow refreshes
of LESS THAN 60 seconds".)
