W3C home > Mailing lists > Public > http-caching-historical@w3.org > February 1996

bypassing and "POST-NO-SIDE-EFFECTS"

From: Paul Leach <paulle@microsoft.com>
Date: Thu, 15 Feb 96 16:39:47 PST
To: dwm@shell.portal.com
Cc: http-caching@pa.dec.com
Message-Id: red-16-msg960216003408MTP[01.52.00]000000b1-96151
I hate it when we have to live with beta mail software... I never 
received my copy of this from the list reflector, so I assume it didn't 
get there.  I sincerely hope its not a duplicate.  I'm only resending 
it because I'd really hate for a needless "POST-NO-SIDE-EFFECTS" 
message to get introduced.
----------
From: Paul Leach
To: "David W. Morris"  <dwm@shell.portal.com>; HTTP Caching Subgroup  
<http-caching@pa.dec.com>
Subject: Re: bypassing
Date: Friday, February 09, 1996 9:40AM

Firewalls are a good point.

In this case, the when the browser knowsa-priori  that the response 
will not be cachable, it can supply  a "Pragma: no-cache" header in its 
request, and the firewall proxy can be smart enough to talk directly to 
the origin server without going up through the proxy hierarchy.

Paul
----------
] From: "David W. Morris"  <dwm@shell.portal.com>
] To: HTTP Caching Subgroup  <http-caching@pa.dec.com>
] Subject: Re: bypassing
] Date: Thursday, February 08, 1996 7:57PM
]
]
]
] On Thu, 8 Feb 1996, Paul Leach wrote:
]
] > A header in the request is superfluous -- the client should  just
] > contact the origin server directly.  E.g.: In the case of POST and '?'
] > for forms -- the form should say whether the result of submitting the
] > form is cachable, and based on that, the browser should either contact
] > its proxy cache, or go diectly to the origin server.
]
] I think he problem we're discusing is what the best method might be for
] the browser and possible intermediate mandatory proxies (e.g., firewalls)
] to learn about possible bypass of caching.
]
] It isn't sufficient for the browser to know since there may be a
] firewall in the path.  Hence, the signal must appear in the data
] stream.  A new method has been suggested. Beefing up the GET body
] in implementations.  Etc.
]
] Dave Morris
] 
Received on Friday, 16 February 1996 00:54:21 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 19:55:57 UTC