- From: Jeffrey Mogul <mogul@pa.dec.com>
- Date: Fri, 18 Jul 97 18:16:07 MDT
- To: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
Unfortunately, the issues list at http://www.w3.org/pub/WWW/Protocols/HTTP/Issues/#MAX-AGE references the wrong email message from Roy as the Proposed Resolution. The reference to http://www.ics.uci.edu/pub/ietf/http/hypermail/1997q1/0650.html should be replaced by a reference to http://www.ics.uci.edu/pub/ietf/http/hypermail/1997q1/0684.html where Roy said "I am changing my proposed change." Anyway, I would like to suggest a few minor, non-normative changes (to Roy's updated proposal): (1) I think this note (in Roy's proposal) Note: An origin server wishing to use a relatively new HTTP cache control feature, such as the "private" directive, on a network that includes older caches which do not understand that feature, will need to combine the new feature with an old Expires value in order to prevent the older caches from caching the response. could be made slightly clearer: Note: An origin server wishing to use a relatively new HTTP cache control feature, such as the "private" directive, on a network including older caches that do not understand that feature, will need to combine the new feature with an Expires field whose value is less than or equal to the Date value. This will prevent older caches from improperly caching the response. (2) I think it would be a good idea to include after the last paragraph of this section (14.9.3): If a cache returns a stale response, either because of a max-stale directive on a request, or because the cache is configured to override the expiration time of a response, the cache MUST attach a Warning header to the stale response, using Warning 10 (Response is stale). the following note: Note: A cache may be configured to return stale responses without validation, but only if this does not conflict with any MUST-level requirements concerning cache validation (e.g., a "must-revalidate" Cache-control directive). In a private email discussion during March, Roy pointed out that this is said elsewhere in the specification. However, I'm concerned that some implementors may misconstrue the discussion in 14.9.3 without such a reminder. -Jeff
Received on Friday, 18 July 1997 18:25:35 UTC