[comp.mail.sendmail] sendmail shock. Bibliography?

barnett@grymoire.crd.ge.com (Bruce Barnett) (04/04/91)

In article <uf3Gv1qg1@cs.psu.edu> schwartz@groucho.cs.psu.edu (Scott Schwartz) writes:


>   My experience has been that Ease is not much better than sendmail,
>   especially considering that the alternatives are so much better.

Ease doesn't give you anything more than sendmail does. Except
an easier to read file, perhaps.

>   The mailer of choice around these parts is Zmailer, by Rayan
>   Zacharaissen <rayan@ai.toronto.edu>.  It is understandable, works
>   well, has readable config files, and it is designed to scale well.

I know little about smail3 or Zmailer. However, I have 246 different
addresses in my mail address test suite that have to be properly
handled. One feature I need is the ability to re-write broken,
brain-damaged addresses. 

Some GE sites have extreme difficulty generating the correct address,
especially when they use UUCP and DECNET to forward mail to our
gateway. I have to re-write the addresses into proper form. This
includes machines providing the wrong domain address.  I also have to
rewrite internet addresses into the proper form when I forward it to
sites with limited delivery agents, like DECNET, UUCP, and SMTP servers.

In other words, instead of telling other sites that their system
should have smarter mail delivery agents, I Make It Work with
Sendmail.

Example: There is one machine who is in domain "XYZ", who insists on
re-writing their address into the form

	user@XYZ.com.user@anything.arpa

I can take that address and write it into the correct *.XYZ.ge.com
address.

Can smail3 or Zmailer do these things? Can anyone give examples?
IMHO, the problem with sendmail is not the syntax (esp. with Ease), it
is understanding EVERY POSSIBLE MAIL ADDRESS VARIATION, and HOW TO
DEAL WITH IT CORRECTLY.

Any delivery agent that can solve those sorts of problems will be
complex. It's not the fault of the MDA, it's because the problem is complex.

Switching to tools other than sendmail will not make the problem go away.
--
Bruce G. Barnett	barnett@crdgw1.ge.com	uunet!crdgw1!barnett