Re: i69: Clarify "Requested Variant" [was: New "200 OK" status codes, PATCH & PROPFIND]

Brian Smith wrote:
> Julian Reschke
>> Brian Smith wrote:
>>> Mark Nottingham wrote:
>>>> PROPOSAL:
>>>> A 201 response MAY contain an ETag response header field 
>>>> indicating the current value of the newly created
>>>> resource's selected representation ETag (i.e., the ETag
>>>> that would be returned if the same selecting headers had
>>>> been sent in a GET request to it).
>>> The above implies that only one resource was created. "the
>>> newly created resource" should become a reference to the
>>> resource pointed to by the Location header. 
>> Yes (with the special case for PUT).
> 
> What do you mean? A PUT can result in multiple resources being created,
> and a 201 response for a PUT should have a Location header. Actually,
> section 10.2.2 of RFC 2616 doesn't say "should" but I think that is the
> intent.

In general, there's no point in sending a Location header for a 201 
response to PUT. If the resource was not created at the Request-URI, the 
request shouldn't have succeeded in the first place.

And, as far as I recall from debugging WebDAV traffic, many servers 
indeed do not include the header.

BR, Julian

Received on Thursday, 14 February 2008 15:07:35 UTC