[Mimedefang] New User

[Kevin A. McGrail]

> I am currently using SpamD instead of letting Mimedefang call
spamassassin.
> Is this how most people use it, or do you let Mimedefang call
spamassassin.
> Is one way better than the other. I am calling spamd/spamc through my
> procmailrc, but if I wanted to let mimedefang take over, what would I need
to do?

Off the top of my head, here's some pros and cons for switching to calling
SA from MD rather than MD passing off to procmail calling SA:

Pros:

1) Lower overhead for each message scanned
2) Procmail scans on a per recipient and mimedefang can scan on a per
message basis

Cons of switching:

1) The Mail::SpamAssassin module doesn't act exactly like the SpamAssassin
daemon or program. I believe there was some code changes in 2.60 to help
this but one of the noticeable problems we just ran into is that you can
ONLY use one .cf file.  SpamAssassin/SpamD will load ALL your .cf files.
Mail::SpamAssassin will just load the first one at least when used with
MIMEdefang (I'm pretty sure from memory that it was a SpamAssassin bug
though).

2) If you are using any per user configuration of any type, the MimeDefang
will need to be coded to mimic that behaviour

3) The MD gets a 1 or a 0 this is spam response and THEN tags it in the
manner it is used to.  If you are familiar with SA, this is a bit awkward.
The first thing I did with MIMEDefang was modify the filter to mimic SA's
old behavior + MD's behavior.  You can see my code at
http://www.peregrinehw.com/downloads/MIMEDefang/mimedefang-filter-KAM


As for your question of how to make the actual switch, I think you might
actually be scanning twice because MD should be using SA if it detected it.
Maybe that is a filter issue though?

Regards,
KAM