Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84 SMI; site sun.uucp
Path: utzoo!watmath!clyde!burl!ulysses!bellcore!decvax!decwrl!sun!guy
From: guy@sun.uucp (Guy Harris)
Newsgroups: net.bugs.uucp
Subject: Re: multihop uucp
Message-ID: <3172@sun.uucp>
Date: Fri, 17-Jan-86 15:22:01 EST
Article-I.D.: sun.3172
Posted: Fri Jan 17 15:22:01 1986
Date-Received: Sun, 19-Jan-86 04:41:11 EST
References: <1564@emory.UUCP> <425@hoptoad.uucp> <647@down.FUN> <1102@ptsfa.UUCP>
Organization: Sun Microsystems, Inc.
Lines: 30

> Having to live with that incompatability every day, I can state that 
> whatever the 'good reason' it is NOT good enough.

If the reason is a security reason (a lot of the changes in Honey Danber
were done for security reasons), it is quite good enough, considering the
number of security holes in vanilla UUCP.  (I'm sure Peter can tell us more
than we wanted to know about this.)

> Note that people upgrading from old System V to Honey-Danber (aka AT&T BNU)
> will have their multi-hop uucp break unless they upgrade all their machines 
> at the same time. Of course, not all vendors offer Honey-Danber uucp so
> --gotcha--.

Not all vendors offer the System V UUCP either, so even if Honey Danber
didn't exist there would be no guarantee that UUCP forwarding would work.

> One hack around it is to ignore multi-hop uucp and write
> a separate forwarder that lives on each machine and which works
> the way mail/rmail works (you would not uucp a!b!c you would then
> "foo a!b!c" in the same way you now "mail a!b!c").

This was such a good idea that Mark Horton hopped into Larry Wall's DeLorean
and invented it several yars ago.  It's called "uusend" and comes with
4.xBSD.  Of course, *it* only works if all the hosts in the path have
"uusend" *and* have their "uux" configured to execute it - and considering
that the V7, 4.1BSD, and System III UUCP's lists of "uux"able commands was
*compiled in*, it may be difficult to reconfigure some hosts to support
it....

	Guy Harris