[Mimedefang] New to MIMEDefang, a question or two

David F. Skoll dfs at roaringpenguin.com
Tue Jul 30 17:30:01 EDT 2002

On Tue, 30 Jul 2002, Jim McCullars wrote:

> Hi, I'm in the process of installing MD and reading all the docs to get a
> handle on how it works.  One thing I'm curious about - the man page says
> that any return value from filter_begin is ignored, and that most of the
> dispositions can be called only from filter - not filter_begin or
> filter_end.  But the sample filter seems to violate this by calling
> action_quarantine_entire_message and action_discard from filter_begin.  Is
> this allowed?

Yes; I need to clarify the man page.  Any action which requires an
$entity argument can only be called from filter (or sometimes
filter_end).  Likewise, any action which is specific to a part (like
action_drop, etc) can only be called from filter.  Actions which affect
the entire message can be called from filter_begin, filter or filter_end.

> My second question is, I can see variables that refer to
> some header values, but is the body of the message readily available to
> the filter script?

Yes; you have to open the file "./INPUTMSG" for input.

> There are some catch phrases that appear in a lot of
> the spam I get that I may want to reject on.  If this is possible, could
> someone provide a small code sample?  Thanks!

if (open(IN, "<./INPUTMSG")) {
	while(<IN>) {
		if (/spam_phrase/i) {
			# handle it

This is not foolproof.  You're better off using SpamAssassin for spam



More information about the MIMEDefang mailing list