W3C home > Mailing lists > Public > ietf-http-wg@w3.org > April to June 2009

RE: Cache-Control private="...", no-cache="..." (was RE: draft-ietf-httpbis-p6-cache-06)

From: Brian Smith <brian@briansmith.org>
Date: Wed, 27 May 2009 09:14:53 -0500
To: "'Yves Lafon'" <ylafon@w3.org>, "'Brian Smith'" <brian@briansmith.org>
Cc: "'Adrien de Croy'" <adrien@qbik.com>, "'HTTP Working Group'" <ietf-http-wg@w3.org>
Message-ID: <004201c9ded5$84f61590$8ee240b0$@org>
Yves Lafon wrote:
> On Tue, 26 May 2009, Brian Smith wrote:
> 
> > I thought I had an explanation of how this could be useful, but my
> > explanation doesn't really jive with what the spec. says. I, too,
> > would really like to hear somebody explain how a cache is supposed
> > to handle private="header1, header2" and
> > no-cache="header1, header2".
> 
> The same way you will handle "Cache-Control: no-cache, max-age=86400".
> It is an error case, and a good implementation will most probably
> choose to take the more restrictive approach (no-cache), or simply
> discard the response for the cache purpose (so treat it as
> uncachable, which is the same as no-cache, but the intent is
> different). Another extreme is to simply reject such responses
> as it is clearly an error.

I was unclear. I was asking, simply, what does
Cache-Control:private="header1, header2" mean? How, exactly, should a cache
process subsequent requests differently than if the cache response had just
Cache-Control:private?

And, how can a cache treat Cache-Control:no-cache="header1, header2"
differently than just Cache-Control:no-cache? How can one cache and
revalidate only part of a response?

Regards,
Brian
Received on Wednesday, 27 May 2009 14:30:53 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:51:03 GMT