Just a few notes as I go.
First, the whole ack/nak stuff seems to be under discussion in the
ietf-ack list and I'd hate to preempt it. I hate also to rely on
communal wisdom rather than looking something up, but I will since this
issue will be resolvable immediately by somebody. I recall that Mark
Crispin had a major objection to some kinds of recursive embeddings.
Would Nathaniel's proposed format for a message rejection create exactly
that kind of recursive embedding?
Nit:
In the above example, the ONLY thing that is not 'boilerplate" is the
choice of boundary string. The phrase "unique-boundary" should be
replaced by a string that does not appear (prefixed by CRLF and two
hyphens) in any of the body parts.
Nope: the string may not appear ANYWHERE in any of the body parts. It
is insufficient that it not appear prefixed by CRLF and two hyphens.
Craig