W3C home > Mailing lists > Public > ietf-http-wg@w3.org > January to March 2008

Re: i22: ETags on PUT responses

From: Henrik Nordstrom <henrik@henriknordstrom.net>
Date: Sat, 05 Jan 2008 22:25:35 +0100
To: Werner Baumann <werner.baumann@onlinehome.de>
Cc: Julian Reschke <julian.reschke@gmx.de>, Mark Nottingham <mnot@mnot.net>, HTTP Working Group <ietf-http-wg@w3.org>
Message-Id: <1199568335.17001.13.camel@henriknordstrom.net>
On fre, 2008-01-04 at 22:14 +0100, Werner Baumann wrote:

> I wonder what the intended use of this Etag, sent in response to a 
> PUT-request, would be. As the entity, sent in a subsequent GET is not 
> guaranteed to be equal to the body of the PUT-request, not a single 
> client knows the entity. But clients get an entity-tag - for an unknown 
> entity. What are they supposed to use it for?

Thr client knows the entity is a direct derivate of the entity in the
PUT. The version of the server resource entity created by the PUT.

It maybe isn't octet equal with the request entity, but for as long as
the strong etag is the same it's the exact version (in octets) created
by the PUT request.

This ignoring any server-side implementation races when dealing with
multiple concurrent PUT requests for the same resource, those are just
server implementation bugs.

This use of ETag is entirely fine for PUT, but I wonder what to do in
case of PATCH... but I guess that octet based PATCH requests should only
be allowed to servers using an octet based store.. it simply gets too
hairy otherwise,


Received on Saturday, 5 January 2008 21:25:47 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 11:10:44 UTC