[rbldnsd] Pre-announce: rbldnsd-0.993p1

[Michael Tokarev]

(..and yes I know most package managers will
sort 0.993p1 AFTER 0.993... i'm sorry for that ;)

Based on recent discussion, I'm releasing something
to test out - rbldnsd-0.993p1, or rbldnsd-0.993 to be.

User-visible changes since 0.992 release are (from
the NEWS file):

  - feature, safety: implement and enforce $MAXRANGE4
    special like this:
      $MAXRANGE /24
      $MAXRANGE 256
    the maximum "size" of a single entry, in number
    of IPv4 addresses it covers.  If an entry covers
    more addresses, it is ignored (and warning is
    logged).  The constraint may be decreased by the
    following $MAXRANGE special, but can not be
    increased.  Global per dataset.

  - feature, safety: ignore incomplete last lines
    (lines w/o end-of-line terminator) in data
    files (to prevent mis-interpreting of incomplete
    data)

  - feature, safety: check for data file changes during
    reloads (while reading data), and abort loading
    (and mark all zones to return SERVFAIL until next
    reload) if a change is detected.

  - safety: do not treat bare numbers as /8 ranges.
      10 -- wrong from now on
      10/8 -- ok
      10-11 -- ok

  - safety: require equal number of octets for x-y
    style ranges:
      1.2.3-2.3.4.5 -- wrong
      1.2.3.0-2.3.4.5 -- ok
      1.2.3.4-2.3.4 -- wrong
      1.2.3.4-2.3.4.5 -- ok
    and the "repeat-last-octet" variant is still
    ok too, obviously:
      1.2-3 -- ok
      1.2.3-4 -- ok
      1.2.3.4-5 -- ok

  - safety: only accept complete, 4-octet IPv4
    addresses in ip4tset, do not allow weird stuff
    like inet_aton() allows:
      10   = 0.0.0.10 -- wrong
      10.1 = 10.0.0.1 -- wrong

  - bugfix: several more small fixes for IP4 address
    parser

  - refine logging a bit, make it less verbose
    (esp. when logging problems)

  - bugfix: query logging (-l) with background
    reloading: the file was not flushed properly
    (resulted in double logging)

  - bugfix: dump (-d) of MX record (generic dataset)
    was incorrect

  - bugfix: wrong subzone in $ORIGIN when dumping (-d)
    combined dataset

  - bugfix: incorect (opposite) evaluation of maxttl

This is quite a long list of (quite minor) changes.

There's no single change in documentation so far,
I will do this part a bit later.  From the above
listing it should be more or less clear what to
expect from this prerelease.  One hint: $MAXRANGE4
may be specified in a separate file listed BEFORE
and data files downloaded from other sources (e.g.
DSBL or CBL or whatever), and it will cover any
data in there, including other $MAXRANGE4s which
may be present in 3rd party data.

Hopefully this (plus, ofcourse, some good pieces of
documentation I mentioned earlier) will help to
eliminate, or at least reduce, problems with corrupted
input.

Rbldnsd homepage is at http://www.corpit.ru/mjt/rbldnsd.html
The NEWS file is at http://www.corpit.ru/mjt/rbldnsd/NEWS
Source is at
  http://www.corpit.ru/mjt/rbldnsd/rbldnsd_0.993p1.tar.gz

Feedback welcome.

/mjt