Patrick O'Keefe wrote:
On Thu, 25 Jun 2009 16:36:28 -0400, Bob Rutledge
<[email protected]> wrote:
Patrick O'Keefe wrote:
...
On z/OS (and I assume other IBM mainframe operating systems) a
leading zero signifies a leading zero, not a switch to octal.
We include leading zeroes in some IP addresses because it
simplifies sorting.
READY
nslookup 10.20.100.4
EZB3170I Server: ...
EZB3172I Address: ...
EZB3170I Name: host...
EZB3172I Address: 10.20.100.4
READY
nslookup 10.020.100.4
EZB3170I Server: ...
EZB3172I Address: ...
EZB3010I *** ... can't find 10.020.100.4: Non-existent domain
...
I forgot that NSLOOKUP uses its own resolver. It implements
different rules than gethostbyaddr(). Try a ping or tracerte or
any application that uses gethostbyaddr() on z/OS (at least 1.8)
and you get the same results with or without leading zeroes.
NSLOOKUP and DIG use their own name resolution schemes.
Interestingly, the Unix command "host" (no TSO equivalent in 1.8)
that I thought used gethostbyaddr() seems to have the octal junk
support. (Why octal is still supported in anything is beyond me.)
My simple-minded answer will be to continue to avoid leading zeroes until the
network folks get consistent. Everywhere.
Bob
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
