I did a quick test to see the differences.
The original has a Return-Path header, the forwarded attachment does not. The forwarded attachment has a Reply-To header that the original did not.
The value of the Return-Path and the Reply-To in the messages are identical but the headers are different. Is this important?
Can anyone point me to something that already exists that may strip out the wrapper?
It would also be helpful to know exactly what headers sa-learn is interested in. Is this documented anywhere or would I need to look through the code and try to figure it out?
I was under the impression that salearn will tokenize any header (even 'Recieved:'), and generally speaking I don't think using a 'resend' feature keeps all the headers. I would second Loren's suggestion to have a script strip out the embedded mime attachment from a message and learn that, that is what I usually do for manual training.
Ryan Moore ---------- Perigee.net Corporation 704-849-8355 (sales) 704-849-8017 (tech) www.perigee.net
