[rbldnsd] Another pre-announce: rbldnsd-0.89p1

[Michael Tokarev]

After some review, I realized that DJB's rbldns just does not
work in current world.  This may sound too strong, but it's true.
DNSBL built on just rbldns will not work.

0.89p1 introduces significant changes (preparation for 0.90 release),
and quite some incompatibilities.  Almost all is due to my attempts
to match BIND's runtime behaviour as close as possible, in both "edge"
cases and normal cases.  It is still very experimental release - I'll
do come code review in a few days.  But I'm asking for help again -
about "proper" behaviour of a DNSBL nameserver.  There are quite a
few strange situations may pop up (e.g., it is not correct to return
NXDOMAIN for 0.0.127.bl.example.com (note incomplete IP), if e.g.
2.0.0.127.bl.example.com exists).

Most significant changes:

  - generic dataset does not understadn SOA RR anymore.  Instead,
    such record may be specified in any data file as $SOA.

  - zone order in command line is important again, at least for now.

  - and, rbldnsd stops searching zones at first matching zone, so
    one should specify all subzones before superzone, or else
    subzones will not be consulted.  I hope to implement some
    workaround here in the future, but there should not be much
    problems with that.

Rest of changes are outlined at http://www.corpit.ru/mjt/rbldnsd/NEWS .

/mjt