On Thu, 5 Dec 1996, Michelle Dick wrote:
> The only change I might suggest for exim is that this be noted in the
> docs a little better (for those folks not up on all the RFCs). When I
> first observed this I was in the dark as to where the Resent-To:
> header was coming from (procmail, smartlist, exim, or my
> smarthost).
Thanks for that suggestion; I have made a note to put in some
clarification for the next edition. In fact, the current chapter 32 not
only needs clarification; it is actually wrong. Sorry about that. The
following comment from the Exim source explains what it actually does.
When there are any "Resent-" headers, it operates on the Resent- set
rather than the normal set.
/* A message is not legal unless it has at least one of "to", "cc", or "bcc".
Note that although the minimal examples in RFC822 show just "to" or "bcc", the
full syntax spec allows "cc" as well.
If we extracted recipients from bcc, then we simply add an empty bcc header,
which is legal according to RFC 822. Do the same for input via SMTP - thus not
disclosing possible bcc recipients. Otherwise we generate to: from the envelope
address fields. Don't bother if there are no recipients (an error case that is
picked up below); such a message isn't going anywhere. */
In other words, it generates "to" or "resent-to" only if called from
another process on the same host, with a list of recipients as its
arguments. This seemed to me to be the best fix for instances like
exim someone@somewhere
Message with missing "to" header
.
but I can see that for a mailing list it may not be good. I didn't think
mailing lists worked this way, as having hundreds or even thousands of
arguments to a command sounds a little dodgy to me. However, clearly it
is being used. I will add an option to cause Exim always to add an empty
bcc: option instead of to: (with resent- if necessary). Will this make
things OK for you? In the meantime, if you can arrange that the input
message already contains an empty resent-bcc: header, Exim should leave
it alone.
--
Philip Hazel University Computing Service,
ph10@??? New Museums Site, Cambridge CB2 3QG,
P.Hazel@??? England. Phone: +44 1223 334714
