1 |
I'm currently running postfix as my mail system. Unlike a recent post which |
2 |
assumed that postfix was overkill compared to sendmail, I tend to feel the |
3 |
opposite, that sendmail is bloated much more than the more simplified |
4 |
postfix. |
5 |
|
6 |
However, when looking for spam blockers, virus checkers, and other message |
7 |
processors, you will typically find multiple sendmail milter implementations |
8 |
as opposed to the postfix filter implementation. |
9 |
|
10 |
Having investigated the milter API and having constructed some milters |
11 |
myself, I currently believe that it should be possible to build a |
12 |
postfix-milter bridge. |
13 |
|
14 |
The milter side of the bridge would support the milter api for compiling |
15 |
sendmail milter-based filters. |
16 |
|
17 |
The postfix side would simply need to wrap a simple smtpd/smtp process that: |
18 |
|
19 |
a) receives mail on the smtpd side. |
20 |
b) pushes the mail through the milter via the bridge api. |
21 |
c) sends the output of the milter back into postfix via the smtp side. |
22 |
|
23 |
Obviously there are some timing issues, etc., to deal with as well as |
24 |
whatever implications there are re: providing a milter-like API for postfix, |
25 |
etc. |
26 |
|
27 |
But assuming these could be resolved, is there general interest in such a |
28 |
beast? Or am I the only one wishing I could integrate tools like MIMEDefang |
29 |
into my postfix email system? |
30 |
|
31 |
I would estimate the building of such a tool to take about 200 hours by |
32 |
leveraging existing smtpd/smtp open source code, and am happy to invest that |
33 |
time if there is some apparent interest in the community for such a tool. |
34 |
I'm just a little leary of investing that amount of my time if there is no |
35 |
real interest in it, just as I would not be willing to invest my time in |
36 |
restoring a Pinto ;-) |
37 |
|
38 |
Dave |
39 |
|
40 |
-- |
41 |
gentoo-user@g.o mailing list |