Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxn!ihnp4!inuxc!pur-ee!pucc-j!purdue!decwrl!glacier!Shasta!ALMQUIST@SU-SCORE.ARPA
From: ALMQUIST@SU-SCORE.ARPA
Newsgroups: net.unix-wizards
Subject: More on previous message
Message-ID: <272@Shasta.ARPA>
Date: Thu, 17-Apr-86 13:41:55 EST
Article-I.D.: Shasta.272
Posted: Thu Apr 17 13:41:55 1986
Date-Received: Sun, 20-Apr-86 17:41:22 EST
Sender: daemon@Shasta.ARPA
Organization: Stanford University
Lines: 26

From: Philip Almquist <ALMQUIST@SU-SCORE.ARPA>

	I learned of one slight lie in my previous message;  the new table
is in use on EPIC, but use on the other LOTS 20's is awaiting the installation
of a new Monitor.

	For DEC-20's, I the new table requires some new magic Monitor
constants.  I think they are NHSTN=NHOSTS*8 and NHOSTN=NHOSTS*3.  The
NICUPD program also needs to have a literal "NIC" (in a GTJFN% string)
changed to an "ARGUS".  Some hack to remove uninteresting hosts (such
as the IBM-PC's and PERQ's at MIT and ISI) might be wise.

	For UNIX systems, the argument to gettable needs to be changed
from "sri-nic" to "argus" and the "netdirprint -nic" stuff needs to be
removed.  Note also the recent messages to mailing lists about
deficiciencies in htable that may cause it to blow up on certain
(non-Stanford) hosts.  Since most UNIX systems get /etc/hosts via
rdist rather than creating it themselves, most UNIX systems don't need
to woory about any of the above.

	I think that the UNIX paragraph also applies to VMS systems.
I don't know about other kinds of systems, though the change should be
trivial: change any references to "SRI-NIC" in the host table
retrieval procedure into references to "ARGUS".
						Philip
-------