procmail
[Top] [All Lists]

Re: Anti spamming

1997-04-10 04:31:00
On Thu, 10 Apr 1997 01:29:21 -0400,
Lars Kellogg-Stedman <lars(_at_)bu(_dot_)edu> wrote:
On Thu, 10 Apr 1997 01:19:55 -0400 (EDT) 
steffi2(_at_)dgs(_dot_)dgsys(_dot_)com (Robert
Nicholson) wrote:
How can you reliably just get the the email address out and leave out
the comments if you are going to use this to do matches?
Right now different user agents send a different From: field and so the
rule is user agent dependent.
A good point.  I hadn't really considered this before, since at the
moment I'm not using *any* recipes that generate autoreplies.  To make
sure you're just getting the email address part of the message, sans

I believe Alan Stebbens wrote up something to cope with this. The
script get-from.rc in the procmail-lib does something like this, only
it actually tries to extract the "human-readable" part instead. 

The Emacs sendmail.el documentation leads one to believe that there
are only three valid formats you can use; I don't recall having seen
many messages which don't fit one of these patterns: 

;;; sendmail.el --- mail sending commands for Emacs.
<...>
(defvar mail-from-style 'angles "\
*Specifies how \"From:\" fields look.

If `nil', they contain just the return address like:
        king(_at_)grassland(_dot_)com
If `parens', they look like:
        king(_at_)grassland(_dot_)com (Elvis Parsley)
If `angles', they look like:
        Elvis Parsley <king(_at_)grassland(_dot_)com>")

Of course, this should not be your authoritative source on this, 
but like I said, most MUAs seem to conform to this. (Hope somebody can
jump up and hit me over the head with an RFC; I'd like that. Honest, I
want to know where it's defined. RFC 822 gives license to, and
examples of, much more contrived variants but I understand they have
been deprecated by some subsequent RFC.)

/* era */

-- 
Defin-i-t-e-ly. Sep-a-r-a-te. Gram-m-a-r.  <http://www.iki.fi/~era/>
 * Enjoy receiving spam? Register at <http://www.iki.fi/~era/spam.html>

<Prev in Thread] Current Thread [Next in Thread>