Re: I-D Action: draft-ietf-httpbis-cache-08.txt

On 6/13/20 9:27 PM, Amos Jeffries wrote:
> On 14/06/20 1:32 am, Ken Murchison wrote:
>> All,
>>
>> I'm trying to verify my understanding of must-revalidate vs unqualified
>> no-cache in responses per the text in Sections 5.2.2.1 and 5.2.2.3
>> respectively.  Other than the point regarding the Authorization header
>> field and must-revalidate, is the difference that unqualified no-cache
>> requires validation for every request and must-revalidate only requires
>> validation after the response has become stale?
>>
> The specification difference is exactly that yes.


Thanks Amos.

> In practice implementations (particularly older RFC2616 based ones) may
> treat "no-cache" as if it were "no-store" and "must-revalidate" as if it
> were unqualified "no-cache".


So, given the possible ambiguity here, would a best practice for 
cacheable resources that should always be validated be for the origin 
server to return maxage: 0 or Expires in the past?

-- 
Kenneth Murchison
Senior Software Developer
Fastmail US LLC

Received on Sunday, 14 June 2020 16:37:41 UTC