[Mimedefang] MIMEDefang 2.55 is released

[Kevin A. McGrail]

I thought about this issue as well and originally considered the same
perspective that you took below before discarding it in favor of the ongoing
position argument that MD is a framework that should be capable of any
milter task that swayed me.

I use a similar argument with SpamAssassin.  SpamAssassin's strength is that
it is an excellent scoring framework for anti-spam.  If I create an
algorithm that rocks for anti-spam (or pro-ham), I can simply integrate it
into the existing framework as SA with an appropriately weighted score.

This framework and reusability is what makes programs like MD and SA
invaluable to me.

I will state though that processing the helo prior to filter_sender would
cause astronomical False Positive rates for my server.  I even know people
using helo processing in filter_end because in their setup, they would have
FPs under their setup.  So I would not suggest a recommended milter
implement filter_helo but otherwise, it's a perfectly valid idea under the
auspices of increasing the flesh on the framework capabilities.

My $0.02,
KAM

> The regular filter sender is hadling the 'helo' quite well and it only
> a command
> away from the 'helo' command...?!? so what is the cost benefit?
>
> 10x David and other for a great tool!