mailfront on IPv6 ?

[John R. Levine]

I've been using mailfront since 1.11. For reasons that are more political
than technical, I want to set up a mail server on IPv6 and see what shows
up. Does anyone have experience using mailfront on v6 connections?

There's an IPv6 set of patches to tcpserver from about 2002 that seems to
work OK, and a TCP session is the same on v6 as it is on v4, but the
environment variables are different.

Also, I have some additional modules that people are welcome to if you
want. There's one for DCC, the distributed bulk counting package, one for
spamassassin that's a little cleverer than Bruce's, one that logs the mail
envelopes in a mysql database, and one that does greylisting, via UDP
queries to a small daemon written in perl.

I slightly modified the CVM plugin to remember the controlling username
after a qmail lookup, and use that in the dcc and spamassassin plugins so
it uses the appropriate per-user config.

Regards,
John Levine, ***@iecc.com, Primary Perpetrator of "The Internet for Dummies",
Please consider the environment before reading this e-mail. http://jl.ly

[Bruce Guenter]

Post by John R. Levine
There's an IPv6 set of patches to tcpserver from about 2002 that seems to
work OK, and a TCP session is the same on v6 as it is on v4, but the
environment variables are different.

All mailfront plugins that would use ucspi-based environment variables
detect the protocol in use and pick up the right prefix. I would still
appreciate hearing what actually happens, though. The add-received
plugin is probably a particularly useful test case.

Post by John R. Levine
Also, I have some additional modules that people are welcome to if you
want. There's one for DCC, the distributed bulk counting package, one for
spamassassin that's a little cleverer than Bruce's, one that logs the mail
envelopes in a mysql database, and one that does greylisting, via UDP
queries to a small daemon written in perl.

I'm definitely interested in DCC and what is more clever in the
SpamAssassin plugin, for inclusion in the main package. The MySQL
logging plugin seems a little too specialized for general inclusion
(though if there is demand I'll add it). Greylisting is interesting to
many too, so I could add that as well, though I don't particularly like
the idea.

Post by John R. Levine
I slightly modified the CVM plugin to remember the controlling username
after a qmail lookup, and use that in the dcc and spamassassin plugins so
it uses the appropriate per-user config.

After a qmail lookup, or a CVM lookup? It sounds useful, though.

[John R. Levine]

Post by Bruce Guenter
I'm definitely interested in DCC and what is more clever in the
SpamAssassin plugin, for inclusion in the main package. The MySQL
logging plugin seems a little too specialized for general inclusion
(though if there is demand I'll add it). Greylisting is interesting to
many too, so I could add that as well, though I don't particularly like
the idea.

I'll send them along.

Post by Bruce Guenter

Post by John R. Levine
I slightly modified the CVM plugin to remember the controlling username
after a qmail lookup, and use that in the dcc and spamassassin plugins so
it uses the appropriate per-user config.

After a qmail lookup, or a CVM lookup? It sounds useful, though.

I use CVM to do a cvm-qmail lookup and remember the CVM_FACT_SYS_USERNAME,
so I can pass it to spamd and dccifd.

Re greylisting, I agree that it shouldn't work, but I can report that I've
been using it for years, and it still makes a fair amount of bot spam go
away.

R's,
John