On October 5, 2014 2:17:28 PM David Jones <djo...@ena.com> wrote:
> Possible extend dkim plugin to bayes ignore header if not dkim signed, > tricky yes, but imho makes sense Why wouldn't all DKIM headers (X-DKIM above and real ones) be excluded? These DKIM headers by themselves are not a good indicator as they can be forged just like any header that is not generated by the local MTA or a trusted relay. Let the DKIM_VALID_AU rule do it's job.
Yep i just writed it to be a change of dkim, but really ment extend bayes to ignore unsigned headers, or maybe better ignore signed headers in bayes ?
Reason i take it up to stop adding 5000 bayes ignore headers to stop abuse of new spamming header to ignore, spammers will always win that battle
