RE: HOW_TO_IDENTIFY_LOCK_OWNER

> From: Jason Crawford [mailto:ccjason@us.ibm.com]
> Sent: Monday, January 28, 2002 8:17 PM
> To: Lisa Dusseault
> Cc: Daniel Brotsky; Clemm, Geoff; Julian Reschke; w3c-dist-auth@w3c.org
> Subject: RE: HOW_TO_IDENTIFY_LOCK_OWNER
>
>
>
> <<
> We can't redefine DAV: owner.  As you suggest, there are current deployed
> uses of DAV:owner and there would indeed be transition issues.
> >>
> If that's the case, let me take a stab at the transition situation of
> Julian's proposal.
>
> Julian proposes a second field for information called
> DAV:lockowner that is
> a child of DAV:lockinfo.  The field is authored by the client and simply
> stored unmodifed by the server.  Julian has defined it in such a way that
> it's clear what "unmodified" means.
>
> So...
>
> Old clients will write to DAV:owner only.  Not to the new field.
> New clients presumably will write to both the old deprecated DAV:owner and
> the new DAV:lockowner.
>
> New servers will (attempt to) preserve both fields.
>
> What will old servers do if they receive a lock info with the new
> DAV:lockowner field but not the old DAV:owner field?  What will
> old servers
> do if the client provides both?   Do some of the interop folks know how
> current "old" servers are coded?

An old server that doesn't ignore the new element would be broken. Is
anybody aware of a server that has this bug?

Received on Monday, 28 January 2002 14:41:16 UTC