RFC 5322, "Internet Message Format", October 2008Source of RFC: IETF - NON WORKING GROUP
Area Assignment: app
Errata ID: 2816
Reported By: Nemo
Date Reported: 2011-05-27
Rejected by: Pete Resnick
Date Rejected: 2011-05-27
Section 3.6.2 says:
from = "From:" mailbox-list CRLF sender = "Sender:" mailbox CRLF reply-to = "Reply-To:" address-list CRLF
It should say:
from = "From:" mailbox-list CRLF sender = "Sender:" mailbox CRLF reply-to = "Reply-To:" mailbox-list CRLF
The text in section 3.6.2 (and everywhere else Reply-To is mentioned) makes it clear that Reply-To is envisioned as referring to one or more mailboxes. But the following chain of productions:
address-list -> address -> group -> DQUOTE DQUOTE ":" ";"
...means that the following Reply-To header would be permitted by the spec:
Reply-To: "" : ;
This header has no mailbox in it at all. So either the reply-to production rule is wrong (and should be a mailbox-list instead of an address-list), or the spec needs to explain what the semantics are for a Reply-To line with no mailboxes in it.
Note also the description of "group" in section 3.4:
The group construct allows the sender to indicate a named group of recipients.
Again, this does not envision using a group in a header like Reply-To. But that is what the current reply-to construct permits.
Summary: The RFC either needs to forbid Reply-To with zero mailboxes, or it needs to explain what the semantics of such a Reply-To are.
Reply-To has allowed address lists (as against mailbox lists) as far back as RFC 733. This would be a change to long held consensus. Not appropriate for an erratum.