W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > October to December 2006

Re: REBIND a parent to a child

From: Geoffrey M Clemm <geoffrey.clemm@us.ibm.com>
Date: Tue, 24 Oct 2006 12:40:32 -0400
To: Julian Reschke <julian.reschke@gmx.de>
Cc: Sriram Malladi <srirampmalladi@gmail.com>, w3c-dist-auth@w3.org, w3c-dist-auth-request@w3.org
Message-ID: <OF20CD49DC.1CAFBA92-ON85257211.005B0663-85257211.005B9A54@us.ibm.com>
I don't think so.  Suppose that /a/b identified a resource R1.  Also 
suppose that /c identified resource R1. 

If you rebind /a to /a/b, that will delete the binding named "b" to R1 in 
the collection named by "/a", so /c will still identify R1, not the folder 
that was identified by "/a", so there is no reason for /c/a to identify 
the collection (unless /c/a originally identified the collection, in which 
case it falls under the criteria of my message below, i.e. that there was 
another binding to /a).


Julian Reschke <julian.reschke@gmx.de> wrote on 10/24/2006 12:28:19 PM:

> Geoffrey M Clemm schrieb:
> > 
> > My response was incomplete.  In case the server does support binding 
> > loops, but /a was the only binding to the collection, so the REBIND 
> > effectively deletes the collection, then the server would fail the 
> > request with DAV:new-binding in the DAV:error status response message 
> > (since the collection does not contain the new binding, since it has 
> > been deleted).
> > 
> > Cheers,
> > Geoff
> Well. But if there's another binding to b, such as /c, all is fine, 
> right? "/a" would be gone, but would be accessible as "/c/a".
> Best regards, Julian
Received on Tuesday, 24 October 2006 16:40:49 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:01:36 UTC