[3021] in Kerberos
Re: long instance names...
daemon@ATHENA.MIT.EDU (Mike_Accetta@cs.cmu.edu)
Tue Feb 22 11:37:58 1994
From: Mike_Accetta@cs.cmu.edu
To: "Roland J. Schemers III" <schemers@SLAPSHOT.STANFORD.EDU>
Cc: Info-AFS-kerberos@TRANSARC.COM, kerberos@MIT.EDU
In-Reply-To: Your message of Fri, 11 Feb 94 16:43:41 -0800.
Date: Mon, 21 Feb 94 16:06:50 EST
Roland,
For the record, we're using long instance names, of course, and they
are sometimes a compatibility problem.
At the time the change made sense since using only the first component
was listed with the BUGS in the Kerberos docs, it was simple to fix,
and we were bootstrapping Kerberos cold in an environment with
pre-existing first component host name conflicts. So far we've been
able to deal with the compatibility problems with source to everything
we use that has Kerberos hooks. If we we're doing it over again, we
would probably try to support both conventions from the outset. We
still might depending on how long it takes V5 to start having an impact
(i.e. how long we continue to depend on V4).
Mike Accetta
