[Top] [All Lists]

Preventing loops?

1999-05-31 20:49:36

I'm currently integrating Sieve into Cyrus.  I'm attempting to prevent
loops by preventing multiple messages with identical message-id's from
being processed twice by a Sieve script.  However, the envelope test
poses certain complications.

A message may be identical to another, but have a different envelope.
For instance, it might be delivered to both 
"leg(_at_)andrew(_dot_)cmu(_dot_)edu" and
"leg+westside(_at_)andrew(_dot_)cmu(_dot_)edu", and I might want to sort based 
on the
envelope address.  Thus, processing one but not the other address may
lead to unfortunate behavior for a user.

I am proposing the following scheme for my implementation, and would
like some feedback:

the sieve script is executed normally upon the first delivery of a

upon subsequent deliveries, the sieve script is run, but only keep,
fileinto, and discard actions are executed; "redirect" and "reject"
are silently ignored.  any duplicate delivery to the same mailbox is

I'd like feedback on this approach...


<Prev in Thread] Current Thread [Next in Thread>
  • Preventing loops?, Lawrence Greenfield <=