On Wed, 11 Dec 1996, Grey Wolf wrote:

> The lookup code would do something like this:
> 
> if (lookup_resolv == LOOK_NET)
>     if (ptr = get_daemon_lookup(req))	/* can't connect?  fall out of loop */
> 	return(ptr);
> return ptr = lookup_in_files(req);
> 
> The upshot of this is that if something gets updated, all you need
> to do is rebuild your daemon.  The time problem is the overhead of
> seeing if the daemon is waiting on that port.  If it isn't, you won't
> be able to connect, it'll drop instantly into the file lookup.

Another advantage of this is that the caemon dould cache lookups for
all clients, whereas a libc implementation could only cache on a
per-client basis.  This is a big win in environments where DNS or
YP are slow.  Slowlaris already does this with nscd, which is
generalized to do the actual lookup and cache results for passwd and
group as well as host information...

--
				Jim Wise
				System Administrator
				GSAPP, Columbia University
				jim@santafe.arch.columbia.edu
				http://www.arch.columbia.edu/~jim
				* Finger for PGP public key *