ecsd(_at_)transbay(_dot_)net wrote:
libspf2 is useless without the proper header files to define symbols used
in the code. The statement on the site "won't compile but this is the
reference version" is irritating if it is 8 months old with no motion.
So, what libspfX should I use?
I have myself been using libspf2 1.0.4, and have been improving and
updating it. I use it at ohmi.org together with an SPF integration into
Sendmail (patch that implements an spf map, much like an access map or
an alias map).
Since libspf2.org seems dead, there's no place to send my patches to, so
I think I will try to take over maintenance of libspf2 v1.0. The API of
the v1.2 version has been changed substantialy, and the state of the
source seems to be mid-work. There are a lot of FIXME comments around,
and I think the missing headers is the tip of the iceberg.
I wrote a scathing thing that said that until Microsoft gets rid of their
licensing (attempts), whatever they offer for protocols should be
scrupulously
ignored. If SPF is taking a hit because "something better has come along",
well, nothing from Microsoft is "better".
[snip]
I think SPF should stay away from this type of comment and
flame-throwing, if it is to be treated as a serious protocol. If
Microsoft has something to contribute that does not threaten the
open-standard/open-source status, I'm willing to consider it.
The fact that they are building Caller-ID on top of SPF is fine, but
that should definitely not affect SPF, which is a well-thought protocol.
Radu.