In
<20020314111429(_dot_)E3752(_at_)melkebalanse(_dot_)gulbrandsen(_dot_)priv(_dot_)no>
Arnt Gulbrandsen <arnt(_at_)gulbrandsen(_dot_)priv(_dot_)no> writes:
ned+ietf-822(_at_)mrochek(_dot_)com
In-reply-to: "Your message dated Tue, 12 Mar 2002 10:10:41 +0000 (GMT)"
<Gsuuxt(_dot_)IC3(_at_)clw(_dot_)cs(_dot_)man(_dot_)ac(_dot_)uk>
Message-id: <01KFA67KP5G80045PS(_at_)mauve(_dot_)mrochek(_dot_)com>
References: <01KF8JCEOCBS0045PS(_at_)mauve(_dot_)mrochek(_dot_)com>
Surely that cannot be correct.
The original message had the normalized header fields
References: <01KF8JCEOCBS0045PS(_at_)mauve(_dot_)mrochek(_dot_)com>
Message-id: <Gsuuxt(_dot_)IC3(_at_)clw(_dot_)cs(_dot_)man(_dot_)ac(_dot_)uk>
and the software I was using at the time copies in-reply-to and references to
references and message-id to in-reply-to, producing
References: <01KF8JCEOCBS0045PS(_at_)mauve(_dot_)mrochek(_dot_)com>
In-reply-to: "Your message dated Tue, 12 Mar 2002 10:10:41 +0000 (GMT)"
<Gsuuxt(_dot_)IC3(_at_)clw(_dot_)cs(_dot_)man(_dot_)ac(_dot_)uk>
Now, RFC 2822 says that the references field SHOULD be constructed using the
message-id and references fields. In effect RFC 2822 chose to interpret the
word "other" in RFC 822 section 4.6.3 (references) as referring to
correspondence other than the current message.
But RFC 2822 was issued more than a decade after this software was written, and
this software interpreted the word "other" in section 4.6.3 as referring to
correspondence other than that what's already listed in section 4.6.2
(in-reply-to). So it doesn't put the message-id field in the references field.
Yes, I would have expected
References: <01KF8JCEOCBS0045PS(_at_)mauve(_dot_)mrochek(_dot_)com>
<Gsuuxt(_dot_)IC3(_at_)clw(_dot_)cs(_dot_)man(_dot_)ac(_dot_)uk>
Only if you're dealing with software that follows RFC 2822. And even then it is
only a SHOULD, not a MUST. Like it or not, the bulk of the world's email
software was written before RFC 2822 was even on the drawing board.
Ned