W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > January to April 1997

RE: Using of Connection header

From: Yaron Goland <yarong@microsoft.com>
Date: Tue, 25 Feb 1997 11:42:23 -0800
Message-Id: <c=US%a=_%p=msft%l=RED-44-MSG-970225194223Z-1661@INET-05-IMC.microsoft.com>
To: 'Jeffrey Mogul' <mogul@pa.dec.com>, "'urbani@ares.mctel.fr'" <urbani@ares.mctel.fr>
Cc: "'http-wg@cuckoo.hpl.hp.com'" <http-wg@cuckoo.hpl.hp.com>
Would the appropriate definition of connection then add "The Connection
header MAY only specify hop-by-hop headers"?
	Yaron

>-----Original Message-----
>From:	Jeffrey Mogul [SMTP:mogul@pa.dec.com]
>Sent:	Tuesday, February 25, 1997 10:29 AM
>To:	urbani@ares.mctel.fr
>Cc:	http-wg@cuckoo.hpl.hp.com
>Subject:	Re: Using of Connection header 
>
>    if the request contains these headers:
>    
>    ---------------------------------------------------------------------
>    Cache-Control: no-cache
>    Connection: Cache-Control
>    ---------------------------------------------------------------------
>    
>    could i use my cache or not ?  Because I must delete de
>    Cache-Control header (because Connection) but I don't know if i
>    must take or not care of it?
>    
>Although the HTTP/1.1 specification does not state this explicitly,
>it is not legal to send
>
>	Connection: Cache-Control
>
>This is because the specification for Connection says
>
>   The Connection general-header field allows the sender to specify
>   options that are desired for that particular connection and MUST NOT
>   be communicated by proxies over further connections.
>
>while the specification for Cache-control says
>
>   Cache directives must be passed through by a proxy or gateway
>   application, regardless of their significance to that application,
>   since the directives may be applicable to all recipients along the
>   request/response chain.
>
>which implicitly prohibits turning Cache-control into a hop-by-hop
>header.
>
>Because it would be against the rules to send this combination
>of headers, the specification does not have to specify a particular
>action to take if someone does it anyway.  (We cannot possibly
>specify an action to take upon receipt of every possible contradictory
>combination of headers.)
>
>However, the robustness principle suggests that it's probably
>safer to ignore the "Connection" header in this case than the
>"Cache-control" header.
>
>-Jeff
>
Received on Tuesday, 25 February 1997 17:20:48 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:32:30 EDT