NACR version 7.0

Wed Jul 28 19:34:01 UTC 1993

    >     It is allowed to submit a block of consecutive Class C addresses 
    >     in a NACR.
    > What about Class B blocks, unlikely though such
    > an event may be?
    This feature is not very useful for Class B as very few 
    sites own blocks of Class B. It is not allowed to avoid
    confusion and to make it easier to detect typos.
Very good reason.

    >     netnum:         198.4.65 - 198.4.79
    > What would be allowed and not allowed here?
    > For example:
    > ....
    All of the above are allowed.
Fine. Makes life easier for script-writers. ;-)

    The rule is that: 
    	o The "firstip" and "lastip" should be separated by 
         	  a "-" (plus zero or more white spaces). 
    	o The trailing 0 (last octet) for the "firstip" or 
    	  "lastip" are optional.
OK, clear.
    	o Also, the first two octets of "firstip" and "lastip"
    	  have to be identical.
Ouch! See below.

    > Also note that your example is a trivial one.
    > I assume/hope you can cope with "overflows":
    >   netnum:         198.4.90-198.5.10
    > too?
    Yes, such NACR's will be rejected by the parser.
I'm afraid I wasn't too clear this time...
What I assumed/hoped was that you would cope
with "overflows" in the sense that they would
be treated properly, not rejected! There are
perfectly valid examples of such "overflowing"
blocks in the RIPE database, which is why my
NACR-generating script accepts them. It didn't
generate "block NACRs" though since I didn't
know it was allowed; instead it breaks blocks
down into separate entries. Given the risk of
rejections on "overflowing" blocks I'll leave
it that way.


	Piet