Re: draft-murchison-webdav-prefer

I think that behavior is acceptable. Please be sure to include the
Preference-Applied response header tho. Just to make it unambiguous.
On Dec 18, 2014 7:26 AM, "Ken Murchison" <murch@andrew.cmu.edu> wrote:

>  All,
>
> I'd like to get some feedback on draft-murchison-webdav-prefer
> <http://tools.ietf.org/html/draft-murchison-webdav-prefer-07> ,
> specifically Section 3
> <http://tools.ietf.org/html/draft-murchison-webdav-prefer-07#section-3>
> (the rest of the document is truly WebDAV specific).  Per a request from
> the Apple calendar client folks, we'd like to extend
> Prefer:return=representation
> <http://tools.ietf.org/html/rfc7240#section-4.2> to apply to a
> conditional PUT request that fails with a 412 (Precondition Failed)
> response.  This eliminates the need for a subsequent GET to fetch the
> current representation of a resource that failed to update because of a
> validator mismatch.  I view this as analogous to Get + If-Range.
>
> Does anyone see any issues with this new behavior?  Does it violate
> RFC7230-7232 in any way?  Are we allowed to extended return=representation
> to failure responses (RFC7240 only discusses success responses)?  Are there
> any other sane interpretations of a 412 response with a
> Preference-Applied:representation header field which would cause ambiguity
> for the client?
>
> Thanks and Happy Holidays!
> Ken
>
> --
> Kenneth Murchison
> Principal Systems Software Engineer
> Carnegie Mellon University
>
>