ietf-nntp Message IDs

Russ Allbery rra at stanford.edu
Sun Mar 16 22:34:26 PST 2003


7. Article posting and retrieval

  I don't think we should refer to RFC 2822 here, or in fact require a
  particular message ID syntax.  NNTP can at least theoretically be used
  for storing any messages with a vaguely mail-like syntax.  I would
  therefore replace the end of this section with:

    Message-ids are, in this specification, opaque identifiers inside
    angle brackets that uniquely identify messages.  They must satisfy the
    following restrictions:

    o  A message-id MUST NOT contain a US-ASCII space, tab, CR, or LF.

    o  A message-id MUST NOT be longer than 250 octets.

    o  A message-id MUST begin with <, end with >, and MUST NOT contain
       any other occurrences of >.

    Additional syntactic restrictions may be specified by the standard
    specifying the format of articles, normally [RFC 1036].  For use in
    NNTP, message-ids MUST satisfy both the restrictions above and the
    restrictions of the relevant article format standard.

    Message-ids are compared octet by octet by the NNTP protocol, and a
    particular octet sequence is considered to be a different message-id
    than any other, differing octet sequence regardless of the semantics
    of that sequence of octets in any other protocol.  As an example, the
    message-ids:

      <"abcd"@example.com>
      <"ab\cd"@example.com>

    are considered distinct by this protocol even though they would be
    considered semantically identical according to the specification in
    [RFC 2822].

  This neatly decouples our document from anyone else's concept of message
  IDs while still stating what NNTP actually requires.  What do people
  think of this approach?  (Is the RFC 2822 example correct?  I think
  Charles had another example that may be better.)

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>



More information about the ietf-nntp mailing list