spf-discuss
[Top] [All Lists]

Re: What about reverse source path?

2004-05-28 18:36:44
On Fri, 2004-05-28 at 18:33, Stuart D. Gathman wrote:

[snip]

I have a socket server in the Python SRS library:

  Thanks, but what I really need is a C example.  First, because I don't
know python.  Second, because I'm of the persuasion that interpreted
languages should not be used for performance critical servers.
  I've already run across a a python CBV implementation that is a socket
map server (http://www.hottis.de/web/verifier/index.html) and a sample
perl socket map server in the contrib directory of the sendmail
8.13.0.Beta2 distribution.  There is also a C implementation (that I'm
actually using) bundled with Cyrus IMAPD 2.2.x, but it's tied rather
tightly to the whole of cyrus.  I.e.: primarily, the super-server
architecture.

My sendmail does not yet support socket server maps,

  It's only in the not yet released 8.13.  The code itself is rather
isolated and easy to port, so I've ported it to 8.12.11-4.6 that is
bundled with Fedora Core 2.

 so I am still
using the program map.

*shudder*  Program maps are evil. ;-)  That's what is used in the
sendmail.cf/SRS integration page referenced on the SPF website.  When I
saw the use of program maps there, it motivated me to find a better way
and it turns out the socket maps are perfectly suited for this kind of
problem.  And possibly for SPF as well.

-- 
-Paul Iadonisi
 Senior System Administrator
 Red Hat Certified Engineer / Local Linux Lobbyist
 Ever see a penguin fly?  --  Try Linux.
 GPL all the way: Sell services, don't lease secrets