[Mimedefang] First two headers not written to INPUTMSG

Sidney Markowitz sidney at sidney.com
Tue Jul 23 14:53:01 EDT 2002


Troy Carpenter <troy at carpenter.cx> wrote:

> I have not tried SpamHaus or SPEWS so I have no data or comment
> on those lists.

You already are checking those because relays.osirusoft.com includes both of
those. The first is indicated by a return of 127.0.0.6 and the latter by
127.0.0.4. I'm not sure from the documentation at osirusoft if 127.0.0.4
contains only the spews list or includes some other manually created lists
too. The reason I am trying those two is that they seem to use criteria that
would minimize "collateral damage", judging from my reading of the
spamhaus.org and spews.org websites. I'm also using the 127.0.0.2 return
from relays.osirusoft.com, which contains tested open relays.

Oh, I just checked the source for the relay_is_blacklisted function. It does
return true for any return value from relays.osirusoft.com, so you are
finding all of those.

Based on your results and a look at the spamcop.net website, I think I'll
add in bl.spamcop.net and see how that does too.

>  I use filter_recipient (as opposed to sendmail or the other Mimedefang
> routines) because it is passed all the important info I use to determine
> a reject: relay, sender and recipient.

I wanted to make it a SpamAssassin test for similar flexibility. The two
differences between your approach and mine are 1) SpamAssassin tests all of
the ip addresses that are in Received headers, not just the first; and 2)
SpamAssassin just adds the results into the total score so it is easier to
incorporate other factors such as your to postmaster example. The downside
is that without the hack I put in to mimedefang.pl SpamAssassin does not see
the final relay, which is the most important one.

 -- sidney





More information about the MIMEDefang mailing list