Subject: Re: UT_HOSTSIZE
To: matthew green <mrg@eterna.com.au>
From: Charles M. Hannum <mycroft@MIT.EDU>
List: current-users
Date: 05/31/1996 06:59:42
matthew green <mrg@eterna.com.au> writes:

> 
> i think an implimentation that knew how to read the current
> format, but, stored new information in a different format
> would be OK.
> 
> eg, it could have a magic number, or anything.  the format
> would include space for any measure of IP address and also
> MAXHOSTNAMELEN chars for the hostname.

This will make the utmp file *huge*.  In addition, you simply can't
predict what the maximum size of an address will be for all protocols.
I would suggest a variable-length format, recognizing that this makes
the implementation slightly more complex.