Update (July 2000): I've flipped to postfix which supports
many of the features that i was adding to smap while giving
me a similar level of security and greatly improved
performance. As such, I'm not planning on any
updates to this code as there are (IMHO) better
alternatives out there. I hope that these patches
have proven useful to folks over the years.
The TIS toolkit is licensed from TIS
They aren't responsible for these enhancements. We, also, don't claim
any liability. They are as-is. Additionally, we expect that people
will use them in adherence to the TIS license restrictions
Stop Spam!
What we did and why?
My friend Bruce and I were trying to
prevent a spammer from reflecting from one of his client's mail
systems. We came up with the following extenions to smap:
This is a work in progress, please check this page for
bugfixes/enhancements, or send them to me (hagan@cih.com).
I'm looking to learn how to beat spam
from you folks in netland.
![[NEW]](../new.gif)
I got two patches which i've not gone
through yet
- Received a patch from Ulrich Eckhardt which fixes a crash when require-full-email=1,check-from-address=1, and the sender
doesn't have a domain name
- Folded in
Ron Snyder'spatches
allowing blocking at a per user/address level
- Enabled runtime changing
off return values, now you can give people whatever error numbers
you want!
- Installed fixes from <
C.T. Nadovich fixing
problems coredumps ruser and domain pointers not
always being initialized
- Working on what appears
to be a bug with having ip addresses in the spam list with 2.0 BETA release,
keep ya posted
- Fixed a stupid on my part that disabled the
anti spam stuff in 2.0 release...sorry
- Flipped to diff format *phew*
- Added ability to zap email addresses (user@domain)
- Fixed a stupid coredump
- Added more patches from
Ulrich Eckhardt (uli@rgw-express.de
- added DNS screening to smtpd
- Put mods into smtpd
for people who like that licence better and/or can't use smap
- Added patches from
Ulrich Eckhardt (uli@rgw-express.de
- BUGFIX: fixed
domain recognition code to try smaller
pieces of the domain name until success/failure, rather
than just barfing on what should have been a good address.
- Ability to screen out spam from both the stated address and actual address
- [in progress] I'm working on adding the ability to do limited
user@site filtering, too
- Concept of internal and external machines and domains
- internal machines may send mail to wherever
- external machines may only send to internal domains (no
reflecting)
- external machines optionally can be restricted to:
- not sending mail appearing from an internal domain
- not sending mail appearing
from a local user (mandate an @ site)
- optionally have a user@site where site has
either a valid MX or A record
- inverses of many of these concepts (local hosts, local domains)
are also in effect
Read the README file!
Still want it?
Older versions, just in case
mindspring's current filter list
Pardon my html, i'm a lynx type hagan@cih.com