Many people have opinioned:
>Leonardo Rodrigues Magalhães a écrit :
>> SA ML, and several others, maintain From address as the original
>> sender of the message, which made me have some troubles whitelisting it.
>>
>> I tought using whitelist_from, but it wouldnt work because there's no
>> single address to whitelist. I have tried whitelist_from_rcvd but,
>> again, there's no single address to whitelist. I ended up doing:
>>
>> whitelist_from_rcvd * apache.org
>>
>
>That's how I do it. if trusted_networks are set correctly, this only
>whitelists mail transmitted by [*.]apache.org to one of your trusted
>servers. In the apache.org case, this should be safe (I think they won't
>relay spam, at least not too many). If the ML uses an ISP/MSP that also
>sends other mail (such as in the case of the postfix ML), this is less
>safe, but you could add other (meta) rules if you want.
>
>Or does someone already have a better way to handle this?
>
>
>> which, until now, seems to work.
>>
>> I would like to know if there's an easier way to whitelist ML that
>> keeps the original sender address as From address of the messages, just
>> like this ML ......
>>
>
>
The whitelist_from_rcvd uses the envelope "From" not the header
"From:". Also, "whitelist_to" is only a few points (6, I think), not one
hundred. But the bayes_ignore is an important suggestion. These work
for me (rDNS is easy to spoof, so *.apache.org is not safe):
whitelist_from_rcvd [EMAIL PROTECTED] mail.apache.org
whitelist_from_rcvd [EMAIL PROTECTED] hermes.apache.org
bayes_ignore_to [email protected]
if more complex expressions were allowed something like:
whitelist_from_rcvd [EMAIL PROTECTED] mail.apache.org
whitelist_from_rcvd [EMAIL PROTECTED] hermes.apache.org
would be preferable.
Paul Shupak
[EMAIL PROTECTED]
P.S. Yes, you could not run these through SA, but the resulting markup
is often helpful in understanding the posting/question.
