A Response with a Set-Cookie on it can be cacheable, if it has a validator
(e.g., ETag or Last-Modified) or freshness directive (e.g., Cache-Control:
max-age, Expires) associated with it.
This is a good thing; some Set-Cookie responses are cacheable. If one isn't,
it should either have no validator or freshness associated with it, or the
appropriate Cache-Control: header (as well as a stale Expires, to catch
HTTP/1.0 clients).
On Mon, Mar 20, 2000 at 11:49:46PM +0300, Michael Ju. Tokarev wrote:
> Just one rationale about -- why I said that squid should not cache both pages.
> Imagine a situation where two users comes to your server in short time.
> Server should attempt to "establish a session" with each one, as you use
> cookie mechainsm. If both users will get the same contents (with headers,
> as squid can cache all or nothing), then your server will think that it is
> _one_ "session". Imagine what can happen here if you implement, say, shopping
> bag at your server!
> With current situation, when at least some pages are cacheable, this is
> possible.
>
> Regards,
> Michael.
>
> P.S. All statements here are just my intention that may be wrong.
-- Mark Nottingham, Senior Developer Akamai Technologies (San Mateo, CA)Received on Mon Mar 20 2000 - 14:30:59 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:52:18 MST