On Wed, 24 Sep 2003, Pat Lashley wrote:
I thought you said above that sieve runs before the 200 result. So arguably, the message hasn't been -completely- recieved by lmtp until sieve is finished.
The message has been recieved, but not accepted.
I did say 'arguably'. I really don't have strong feelings about whether lmtpd makes its Received header available to sieve or not. The main issue is the Return-Path; which should not (visibly) exist until the message is placed into the mailbox.
Sieve implementations are not required to implement the envelope extention.
I didn't mean to imply that they were. But a procmail-type interface makes it unreasonably difficult to do so cleanly.
Local delivery agents similar to procmail have been used for quite some time. They're certainly not high performance, but they're far from obsolete.
Again, I didn't mean to imply that they were obsolete; only that better interfaces to that functionality have evolved. I would base any stand-alone filtering system on the newer mechanisms rather than hanging them on the old .forward system.
-Pat
