RE: Marshalling Depth > 0 responses for REPORTs, WAS: Replacing t he Label header with a DAV:labeled-version report

   From: Julian Reschke [mailto:[email protected]]

   > From: Clemm, Geoff

   > RFC3253 states that the marshalling is in the DAV:prop node, and
   > we should not change that unless there is some significant problem
   > with that marshalling.

   RFC3253 states that for depth != 0, the response must conform to
   whatever RFC2518 has to say about Multi-Status response bodies.
   ...
   I think by now we can safely assume that we won't get any kind of
   interoperability for these kinds of reports unless we clarify the
   format.

I agree that we need to clarify what is intended by 3253 for Depth>0
REPORT response bodies.  In particular, RFC2518 says that a DAV:prop
node should only contain property values, while RFC3253 says that
the DAV:prop node in a Depth>0 REPORT response should contain a
report.

   I haven't seen any server or client implementing the current
   RFC3253 format yet, so I think clarifying *plus* choosing a more
   logical format makes a lot of sense.

I believe we have two alternatives being proposed:

(a) Clarify that the DAV:prop node contains reports instead of
properties, when used in a DAV:response node in a top-level
DAV:multi-status response body for a Depth>0 REPORT.

(b) Change the marshalling defined by RFC3253 so that the report
appears in a DAV:report node instead of the DAV:prop node.

The advantage of (a) is that it is consistent with what appears in
RFC-3253.  This means that there is at least some chance that someone
implementing RFC-3253 would pick this alternative.

The advantage of (b) is that it is more consistent with RFC-2518,
but unfortunately is not consistent with RFC-3253 and is not something
that anyone would implement based on existing RFC documents (i.e. there
is no way they could guess to use a DAV:report node).

Clearly, Julian prefers (b).  While I prefer (a), I'd be happy to
go with (b) if that is the preference of the working group.

Anyone else want to chime in here?  In either case, I will add this
issue to the ERRATA document.

Cheers,
Geoff