Re: Associating HTTP headers and bodies

On Thu, Dec 24, 2009 at 8:55 AM, Florent Georges <fgeorges@fgeorges.org> wrote:
>  Hi,
>
>  In p:http-request, a multipart request or response content is
> represented as:
>
>    <c:multipart
>      content-type? = string
>      boundary = string>
>        c:header*,
>        c:body+
>    </c:multipart>
>
>  While this maps directly the HTTP entity content structure,
> this fails to express the semantics: a multipart is a set of
> parts, each of them having a set of headers and a content:
>
>    <c:multipart
>      content-type? = string
>      boundary = string>
>        c:part+
>    </c:multipart>
>
>    <c:part>
>        c:header*,
>        c:body+
>    </c:part>
>
>  This is of course equivalent, but in the later case the
> physical structure matches the logical structure.  Furthermore
> (and I would say more importantly), the former case requires
> grouping to get the headers corresponding to one part, and I am
> afraid that will be blocking for same applications.

If you look at section 6.1 of RFC 2387, you'll see that body parts
only require Content-* headers.  As of the current draft, that is all
we support and it is sufficient for transmitting multipart/related content.

I'd be interested in any RFC that describes a header that isn't a
content header as described in RFC 2387 that is intended to be
used in a multipart entity body rather than as a header for the
multipart/related message itself.

In the future, we can support multipart bodies with arbitrary headers, as
you  suggest with a separate element, but we haven't found any
requirement to do so as of yet.

[1] http://tools.ietf.org/html/rfc2387

-- 
--Alex Milowski
"The excellence of grammar as a guide is proportional to the paucity of the
inflexions, i.e. to the degree of analysis effected by the language
considered."

Bertrand Russell in a footnote of Principles of Mathematics

Received on Thursday, 7 January 2010 18:41:19 UTC