From: Aquarion (nicholas@aquarionics.com)
Date: Mon Sep 02 2002 - 19:39:00 BST
On Mon, Sep 02, 2002 at 04:24:14PM +0100, Stuart Langridge wrote:
>
> Aquarion spoo'd forth:
> >
> > On Mon, Sep 02, 2002 at 02:05:11PM +0100, Stuart Langridge wrote:
> >>
> >> Stuart Langridge spoo'd forth:
> >> > 4. If it doesn't find one, or the response of the request doesn't seem
> >> > to support it, then die (see note 1 below).
> >>
> >> The way I implemented this is to reject all "referring" urls that do
> >> not have one of the following content-types (from the HTTP headers);
> >> text/xml, text/xhtml+xml, text/html
> >>
> >> Should there be any others in this list?
> >
> > shockwave? :-)
>
> Nah mate ;)
>
> > Check on text/*?
>
> Hm. Could do that. Are there any flaws with doing that? I'd much rather
> that stuff was a proper HTML content-type, because then it's more
> comfortable for me to make the assumption that it *is* HTML and I can
> therefore blithely parse a title out of it :)
>
> (Memo to self: check what happens if someone tries a titleless page)
>
> > Oh, and beware compressed pages.
>
> Yeah. Hm. I hadn't thought of that. Then again, are there servers that
> serve pages with Content-Encoding: gzip whether you ask for it or not?
> I thought you had to request it in your request, since you can't expect
> that J. Random Browser will support it...?
Nope. Until recently (Recently being defined as "My Girlfriend
complaining it was crashing IE4") Apache compressed every page
Aquarionics sent. Speed increase over dialup was noticable...
Oh, and could you make your pingback thing produce proper error
messages, so I can work out why my pingback request is failing? :-)
-- Aq
This archive was generated by hypermail 2.1.5 : Mon Sep 02 2002 - 21:05:00 BST