On Saturday 11 November 2006 16:00, John A. Martin wrote:
"Scott" == Scott Kitterman
"Re: Re: Apache SpamAssassin SPF checks"
Sat, 11 Nov 2006 15:30:42 -0500
Scott> On Saturday 11 November 2006 14:40, John A. Martin wrote:
>> 2. When PREPEND is used in Recipient Restrictions the
>> pertinent PREPEND will be applied to the same message once for
>> each recipient, ie. multiple identical PREPENDs appear for
>> multiple recipients including "blind" recipients.
Scott> I had a similar problem with the message quota limit
Scott> function I implemented in my MSA oriented derivative of
Scott> tumgreyspf. Until I took this into account I was counting
Scott> addressess and not messages (didn't work so well).
Scott> What I did was save the queue_id and only run the quota
Scott> routine once per queue_id. When I do the Prepend patch for
Scott> tumgreyspf, I'll probably do something like that to solve
Scott> this problem.
At least for the policy daemon ISTM that the queue_id should be saved
in a disk cache because IIRC there is no guarantee that all recipients
of one queue_id would be processed by the same instance of the daemon.
The notion of managing such a cache made me set aside for a while such
a modification to the Postfix policy-spf.pl.
I do write it to disk. My current approach is certainly overly simplistic for
a high traffic site. I've you've got design hints for the disk cache, I'd
appreciate it. I can do it in Python for tumgreyspf and let you know how it
works out.
Thanks,
Scott K
-------
Sender Policy Framework: http://www.openspf.org/
Archives at http://archives.listbox.com/spf-discuss/current/
To unsubscribe, change your address, or temporarily deactivate your
subscription,
please go to http://v2.listbox.com/member/?list_id=735