I would say it is one way to solve the issue caused by checking mail midstream but that MDs solution is a bit more robust. I use it myself and similar routines because I use spamc called from MD to allow me to run spamd wherever I have cycles and ram. Regards, KAM
Henrik K <h...@hege.li> wrote: >On Wed, Aug 14, 2013 at 08:18:55AM -0400, David F. Skoll wrote: >> On Wed, 14 Aug 2013 14:10:55 +0200 >> Axb <axb.li...@gmail.com> wrote: >> >> > isn't Return-Path added my MDA? seems to me this rule will only >work >> > on systems which run SA after delivery, and not in "gateway mode". >> >> Not necessarily. We use it in gateway mode with MIMEDefang, which >> synthesizes a Return-Path: header as well as the first Received: >header >> before passing the message to SpamAssassin. >> >> If other gateway-mode milters or any other gateway-mode mechanisms >don't >> do this, then yes... the rule won't work. > >Correct way is to use EnvelopeFrom pseudo-header as per documentation.. > >http://spamassassin.apache.org/full/3.3.x/doc/Mail_SpamAssassin_Conf.html#envelope_sender_header_name_of_header
