MS 021c: Content-Type: multipart/related;charset clarification, please

[Gait Boxman <gait.boxman@tie.nl>]

Hi,

can someone clarify me on section 7.2.1 of Messaging Service Specification V.021c?

It says that the Content-Type header for the multipart has a charset attribute, and I'm puzzled on it's use.

Why would there be a charset specification on a multipart which can contain multiple bodyparts using different encodings and charsets?

I could think of only two reasons:

1. to specify an alternate character set for the multipart header and boundary data,

2. to specify a default character set for the contained body parts.

However, RFC 2046 states (on the bottom of page 18) that all header fields and boundary delimiters are represented in 7bit US-ASCII, which rules out reason one. Furthermore, RFC 2046 states (on the bottom of page 17/top of page 18) a the default RFC 822 content-type for bodyparts with no headers.

So far, I have found that the charset attribute is applicable for text content types, but not for multiparts.

Did I miss something in the RFC's, is there some other source I did not check. Or was it put in by accident?

 

Thanks, Gait Boxman.