Subject: Re: Hesiod thoughts
To: None <current-users@NetBSD.ORG>
From: der Mouse <mouse@Collatz.McRCIM.McGill.EDU>
List: current-users
Date: 10/30/1995 08:35:02
> Isn't getpwnam() and company part of SPEC 1170?

Nobody is proposing to do away with getpwnam(), or the rest of the
keyed-lookup routines.  Just getpwent().

> By ditching it aren't we trashing the ability of NetBSD to be
> standards conformant?

People put far too much store in standards conformance for the sake of
standards conformance.  Before blindly conforming to a standard, we
should first decide whether the standard in question is worth
conforming to.

> I think something like a database service switch similar to
> nsswitch.conf in Solaris is the better way to go for ALL services.

What "ALL services" does this apply to?  Passwd database lookups, at
least, presumably.  Hostname->address lookup?  Address->hostname
lookup?  Service name lookup (a la /etc/services)?  Timezone name to
rules mapping?

> Work on this is already being done, why not accelerate that work
> rather than removing functions that don't need to be removed?

You seem to have missed the point: Hesiod makes it difficult to
implement getpwent(), impossible without some kludge like the
passwd-<n> CNAMEs or zone transfers.  Other perfectly reasonable
database services do likewise; imagine what the DNS would have broken
if people used gethostent() the way they use getpwent().  That, not the
absence of some nsswitch-like mechanism, is why we're talking about
deprecating gethostent().

					der Mouse