On 7/26/23 2:34 AM, Benny Pedersen wrote:
milters should not be spam scanners, spamassassin is better
{spamass-milter,milter-spamc} combined with SpamAssassin cause me to
question the veracity of that statement.
Milter implies doing the filtering during the SMTP transaction. I
consider the ability to reject messages that SpamAssassin declares as
(bad enough) spam at SMTP time to be a good thing.
maybe use bind9 rpz to change spf data for stupid domains, mostly
freemail domains in this category ? :)
Will you provide a description of what you would have an RPZ do?
I know that you can specify local data overriding the upstream DNS zone
data.
All of the local data would be locally and artificially baked. I guess
you could do some sort of query and create a fresh RPZ record with a
different "all" value. Or are you thinking something else?
This might be a use case for RPS to more dynamically alter the record.
But I've not messed with RPS much. Conversely I've done a lot with RPZ.
Grant. . . .
