RE: PATCH vs multipart/byteranges vs Content-Range

Julian Reschke wrote:
> Brian Smith wrote:
> > Actually, this is wishful thinking on my part. HTTPbis should have 
> > that requirement that Content-* cannot be ignored for any 
> > method that uses a request body (especially POST), but RFC 2616
> > only has the requirement for PUT. Can a new issue be created for
> > that?
> 
> Why would that have anything to do with the presence of a 
> request body?

A server processing a GET or DELETE can ignore all Content-* headers in
the request, because the the request body doesn't have any semantics for
GET and DELETE, and Content-* describe the request's entity.

POST, PUT, and PATCH cannot ignore Content-* headers in the request
because those Content-* headers are needed to understand the request
body enough to process it--especially Content-Range and
Content-Encoding.

Another issue is "What does Content-*" mean? Does it mean all headers
prefixed with "Content-", or just those headers prefixed by "Content-"
that are defined in RFC 2616?

- Brian