Where Bennett had,
T> * $^To:.*$LOGNAME
Dallman asked,
R> Would this not be better as follows?
R> * $ ^TO$LOGNAME
No, and you answered your own question, Dallman:
R> Or, if not, would it be because the object is to restrict
R> the parse to mail actually sent to $LOGNAME and not cc'd, etc.?
Exactly. A vacation response should be sent only in reaction to mail that is
To: the vacationer. If the address of the vacationer does not appear, or it
is only in Cc:, Apparently-To:, Resent-anything:, a visible Bcc:, nor (in the
absence of To:) Received:...for or Delivered-To:, those don't count.
T> * !^List-
T> * !^Mailing-List:
T> * !^X-.*-List:
T> * !^X-Listprocessor-Version:
R> That's a lot of "List" stuff that maybe could just be handled in one
R> line? Certainly the last makes the third redundant, anyway.
No, the fourth does not include the third, nor vice versa. The second and
the third could be combined, though:
* !^(X-.*|Mailing)-List:
R> So how about something like:
R> * ! List
R> and be done with it?
Way too coarse. What if I send you a personal message, not knowing you had
just left on vacation, and the subject is "Are you as sick of the procmail
list as I am?" or "I'm going to meet Calista Flockhart"? What if you were
getting personal mail from somene named Alistair?