W3C home > Mailing lists > Public > ietf-http-wg@w3.org > April to June 2011

Re: #290 [was: SHOULD-level requirements in p6-caching]

From: Mark Nottingham <mnot@mnot.net>
Date: Thu, 5 May 2011 16:15:33 +1000
Cc: "Roy T. Fielding" <fielding@gbiv.com>, Poul-Henning Kamp <phk@phk.freebsd.dk>, HTTP Working Group <ietf-http-wg@w3.org>
Message-Id: <99BA74E9-0407-47A2-A802-350AD228DA81@mnot.net>
To: Adrien de Croy <adrien@qbik.com>

It sounds like you're suggesting changing

> When the no-cache pragma directive is present in a request, a cache MUST treat the request as if it contains the no-cache cache-control directive [ref], unless the Cache-Control header is also present and the cache obeys those cache-control directives [ref] that it contains (excepting any extensions that are not understood [ref]).

to 

"""
When the no-cache pragma directive is present in a request, a cache MUST treat the request as if it contains the no-cache cache-control directive [ref], unless the Cache-Control header is also present.
"""

in the proposal below.

Roy, are you OK with dropping the 'obey' language?



On 05/05/2011, at 1:57 PM, Adrien de Croy wrote:

> 
> having the override be conditional on whether the CC header contained understood directives or not seems like not a great idea to me.
> 
> I'd have presumed that if the CC header exists, Pragma should be ignored - regardless of what directives are contained or whether they are understood or enacted.
> 
> The reasoning being:
> 
> * Pragma was an HTTP/1.0 header
> * anyone implementing this RFC is implementing HTTP/1.1
> * Pragma was recommended in RFC2616 to cope with HTTP/1.0 caches
> 
> therefore an HTTP/1.1 client is only sending Pragma to target HTTP/1.0 intermediaries, which implementations of this spec are not. The only other case where an intermediary would expect to see a Pragma is from an HTTP/1.0 client, in which case it wouldn't expect to see C-C.  HTTP/1.1 clients are still perfectly free to send whatever C-C directives they choose.  It's actually more deterministic to ignore Pragma than treat it as a repeat C-C header (and therefore combine the headers).  Especially since precedence of cache directives isn't explicitly defined (another issue).
> 
> Regards
> 
> Adrien
> 
> 
> 
> 
> On 5/05/2011 3:04 p.m., Mark Nottingham wrote:
>> On 05/05/2011, at 12:32 PM, Roy T. Fielding wrote:
>> 
>>>> """
>>>> When the no-cache pragma directive is present in a request, a cache MUST treat the request as if it contains the no-cache cache-control directive [ref], unless the Cache-Control header is also present in the request.
>>>> 
>>>> A client SHOULD include both pragma and cache-control directives when sending a no-cache request , unless Cache-Control: no-cache is omitted to target other Cache-Control directives at HTTP/1.1 caches, while precluding HTTP/1.0 implementations from using a cached response.
>>>> """
>>>> 
>>>> Is that workable?
>>> Sounds like torture to me.  What is wrong with just ignoring Pragma
>>> if Cache-Control is given and obeyed?
>> 
>> "Obeyed" really needs to be qualified to those directives that are understood. I've have a go below; if you have suggestions, please provide specific text.
>> 
>> """
>> 3.4.  Pragma
>> 
>> The "Pragma" header field allows backwards compatibility with HTTP/1.0 caches, so that clients can specify a "no-cache" request that they will understand (as Cache-Control was not defined until HTTP/1.1). When the Cache-Control header is also present and understood in a message, Pragma is ignored.
>> 
>> In HTTP/1.0, Pragma was defined as an extensible field for implementation-specified directives for recipients. This specification deprecates such extensions to improve interoperability.
>> 
>>      Pragma           = 1#pragma-directive
>>      pragma-directive = "no-cache" / extension-pragma
>>      extension-pragma = token [ "=" ( token / quoted-string ) ]
>> 
>> When the no-cache pragma directive is present in a request, a cache MUST treat the request as if it contains the no-cache cache-control directive [ref], unless the Cache-Control header is also present and the cache obeys those cache-control directives [ref] that it contains (excepting any extensions that are not understood [ref]).
>> 
>> A client SHOULD include both pragma and cache-control directives when sending a no-cache request, unless Cache-Control: no-cache is purposefully omitted to target other Cache-Control response directives at HTTP/1.1 caches. For example:
>> 
>> GET / HTTP/1.1
>> Host: www.example.com
>> Cache-Control: max-age=30
>> Pragma: no-cache
>> 
>> will constrain HTTP/1.1 caches to serve a response no older than 30 seconds, while precluding implementations that do not understand Cache-Control from serving a cached response.
>> """
>> 
>> --
>> Mark Nottingham   http://www.mnot.net/
>> 
>> 
>> 
>> 
> 
> -- 
> Adrien de Croy - WinGate Proxy Server - http://www.wingate.com
> 

--
Mark Nottingham   http://www.mnot.net/
Received on Thursday, 5 May 2011 06:17:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:51:40 GMT