Matt Kettler wrote:
In "old practice", 10.1=10.0.0.1 (a.b = 256^3 * a + b), and not
10.1.0.0.
Really? That's a new one by me.
This is a (deprecated?) BSD practice. it was coded in inet_addr.c and
was thus imported by many systems.
# ping 127.1
PING localhost (127.0.0.1): 56 data bytes
64 bytes from 127.0.0.1: icmp_seq=0 ttl=255 time=0.144 ms
...
# traceroute 10.1.2
traceroute to 10.1.2 (10.1.0.2), 64 hops max, 40 byte packets
...
# traceroute 10.678
traceroute to 10.678 (10.0.2.166), 64 hops max, 40 byte packets
...
These work at least on *BSD and windows (replace traceroute with tracert
of course).
Since the advent of ipv6 and rewrite of inet_aton things using
getaddrinfo(), this seems deprecated. so:
#uname
NetBSD
# telnet 10.678
10.678: No address associated with hostname
See
http://article.gmane.org/gmane.comp.db.postgresql.devel.general/27821
(In particular, see "Paul Vixie" words).
However, SA should perhaps fall in-line with how sendmail does this. In sendmail
access configurations you don't need the trailing . (ie: 10.1 would imply
10.1.0.0/16 not 10.1.0.0/32)
http://www.sendmail.org/m4/anti_spam.html
Point taken. It's safe to agree with widely deployed MTAs.
