I did some serious testing/programming today, and the results are quite
encouraging. :) I already have half an SPF socket map daemon working.
Through fault entirely mine, we have some duplication of effort. Let me
Several weeks ago I started working on an SPF map for sendmail as well.
The development work is done, and I am using/testing the results on my
mailserver with great results. Some of the vitals of this project are:
- SPF2MAP is a sendmail map, compiled in.
- I've fully integrated all needed FEATURE() m4's
- The sendmail source RPM is available
(sendmail-8.13.1-2.1spf2.src.rpm), compiles and installs cleanly on
Redhat 9 (the only system I have access to).
- The binary RPM's for RH9 are also available, and require
libspf2 >= 1.0.4 and < 1.2.0 installed
- Support for both lisbpf2 1.0.4 and 1.0.6 (avail at ohmi, includes
- support for WANTSPF access map (ie list of domains that can be
rejected for not publishing SPF records). By default the feature is
enabled, but the included list (/etc/mail/wantspf) is empty, so no
domains are rejected for non-publishment.
I spent a great deal of time with writing and testing m4 files, in order
to make it easy for 100% of sendmail admins to install SPF checking on
I am hoping that the resulting patch is in tune with the sendmail
philosophy of implementing such features. I tried hard to make the
integration thorough and consistent with the other sendmail
configuration features. My goal is to have sendmail adopt this feature
and build it into their distribution, so that deployment of SPF can gain
momentum with a minimum of headaches from mail admins.
I will shortly start working on an integrated patch for SES.
I will shortly publish the download link for the sendmail RPM and