- From: Geoffrey M Clemm <geoffrey.clemm@us.ibm.com>
- Date: Tue, 25 Nov 2003 09:02:20 -0500
- To: "'Webdav WG'" <w3c-dist-auth@w3c.org>
- Message-ID: <OF69D8148A.583D3988-ON85256DE9.004C52A1-85256DE9.004D1A45@us.ibm.com>
I'm not sure what the benefit would be of this additional error code for MOVE. If the client wants to try COPY/DELETE as an alternative to MOVE, then it should do so. I don't see that the clients decision to do so would be affected by a special error code here. Cheers, Geoff Julian wrote on 11/24/2003 02:31:33 PM: > Helge Hess wrote: > > > the specification is a bit unclear on what we are supposed to return in > > the case that a MOVE cannot be completed because request URL and > > destination URL are on different servers or are on different > > subnamespaces on a single server. > > I would expect some return value that signals to the client that the > > URLs are on different "drives". The client could then try to implement > > the MOVE logic on the client side (eg using PUT to dest then DELETE in > > src). > > Any suggestions? > > Right now, RFC2518 allows both -- rejecting the request, or doing a > "best effort" approach on the server. > > The BIND draft (<http://wwww.webdav.org/bind>) clarifies that a client > that really wants to move resources (keeping all live properties and > resource-ids) can do a REBIND method call that is guaranteed to fail if > a "real" move operation is impossible. > > If a server wants to reject MOVE, that's certainly possible although > existing clients may not be handling this very well. A 403 (Forbidden) > seems to be the right status code in this case. It would be nice if > there'd also be a precondition code for this case (Geoff, could we add > that to BIND as additional MOVE semantics?).
Received on Tuesday, 25 November 2003 09:04:26 UTC