From: Ned Freed
As for updating the specification, I have no problem with adding
>> a statement that the result of an MX lookup MUST NOT be a
>> CNAME if there is consensus to do so. I strongly object,
however, to adding statements along the lines of "this document
>> doesn't update this other RFC". That's a path to madness.
Trevor Paquette wrote:
I really like this idea. This will solve the problem of
misinterpretation of the RFC.
I doubt that because this issue predated RFC 2821.
While hugely successful over the years, one of the problems with the
style of RFC documents (by design) is that they tend to serve a wide
range of audiences. Specifically, they tend to blend both functional
design specifications with technical design specifications. The former
typically serves management, operations, possible marketing. The latter
typically serves development and implementation details.
So it depends on who is reading the RFC.
That said, I do not advocate any change to RFC 2821bis, but if there is
something to point out, IMO, it should be addressed to two parties:
- DNS operators MUST NOT add CNAME records in MX records as this
can create extra unnecessary queries.
- SMTP DNS resolvers SHOULD be prepared for MX lookups
resolving to CNAME records.
This targets both audiences.
Hector Santos, CTO