On Mon, 01 May 2006 19:17:27 +0200, Mark Nottingham <[EMAIL PROTECTED]>
wrote:
[...]
Well, another way to set these headers is letting the author do it. So
I can see why there's a requirement on UAs here from that perspective...
I was thinking of something like
"UAs must not allow the Content-Length, Connection or Keep-Alive headers
to be overridden."
instead of
"UAs must set the Connection and Keep-Alive headers as described by the
HTTP specification, and must not allow those headers to be overridden."
because HTTP already tells UAs how to deal with these headers, and they
don't always have to set them.
Same sort of text for the Host header.
Given that we now have a separate thread on headers and everything else in
this list of comments has been addressed I consider this thread as closed.
Thanks for your comments!
Also, add TE to the list of headers that UAs MAY add.
It MUST NOT be overriden?
Yes.
That's been done.
--
Anne van Kesteren
<http://annevankesteren.nl/>
<http://www.opera.com/>
