W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > January to March 2004

Re: redirect references protocol: automatic update

From: Geoffrey M Clemm <geoffrey.clemm@us.ibm.com>
Date: Fri, 9 Jan 2004 07:30:33 -0500
To: webdav <w3c-dist-auth@w3.org>
Message-ID: <OFFD455C38.EABDD7C2-ON85256E16.004440D0-85256E16.0044AF6C@us.ibm.com>

One possibility is to have an additional argument to MKREDIRECTREF,
e.g. DAV:stable, that means that the client forbids the server from
auto-updating the reference.  A client that doesn't care would leave
off this parameter, and the server would do what it wanted, while a
client that cared would specify this parameter, and a server that doesn't
support stable redirectref's would fail the request.

Cheers,
Geoff

Julian wrote on 01/07/2004 06:43:30 PM:

> 
> Geoffrey M Clemm wrote:
> > I don't see why we couldn't require this (redirects created by some 
other
> > mechanism don't have to act the same way as redirects created by 
> > MKREDIRECTREF.
> > 
> > So we could/should say that the value of a redirectref changes only 
when
> > it is explicitly updated by a PROPPATCH of the DAV:reftarget property,
> > and not as a side-effect of some other operation such as MOVE.
> 
> That's exactly what I *don't* want to say, because in reality there are 
> servers that have a concept of redirect reference resources, *although* 
> they auto-update their target. This is completely consistent with 
> RFC2616, and I don't see why the redirect spec can or should *forbid* 
> that. I do agree that this behaviour may be confusing, and that it makes 

> sense to discourage it, but we can't change what's already there...
> 
> Julian
> 
> 
> -- 
> <green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760
> 
Received on Friday, 9 January 2004 07:30:43 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:17:51 UTC