Path: utzoo!mnetor!uunet!husc6!think!ames!pasteur!ucbvax!ucdavis!ccdan
From: ccdan@ucdavis.ucdavis.edu (Dan Dorough)
Newsgroups: comp.mail.misc
Subject: Re: operator precedence
Message-ID: <1533@ucdavis.ucdavis.edu>
Date: 30 Mar 88 09:46:17 GMT
References: <23560@hi.unm.edu>
Reply-To: ccdan@ucdavis.ucdavis.edu (Dan Dorough)
Organization: University of California, Davis
Lines: 30
Keywords: '!' and '%'

In article <23560@hi.unm.edu> kurt@hi.unm.edu (Kurt Zeilenga) writes:
> What do you do with something like:
> 	a!b%c		(where a, b, and c could domainized)
> 
> I prefer:
> 	a!b@c		(deliver to c)
> because many folks here address things to a!b and I will forward them
> to d for forwarding to c.  So I give d a "a!b%c@d".   Now, I could give
> d "a!b@c@d", but that has two '@', or just "a!b@c".
> 
> What is considered defacto standard way(s) of rewriting "a!b%c"?

the '%' hack isn't actually documented anywhere that i
know of.  for this reason, it could be considered the
"least binding token".  it's been my experience that it
is far better to resolve
	host!left%right
on the '!' and thus send <left%right> to "host" for delivery.
the advantage of the % hack is that it allows for mail
to pick it's way through systems that will hopefully
leave the '%' parts alone until the the very last.  if you
start giving it higher precedence than '!', the results
will be unpredictable.  the only time you would resolve on
the '%' in an address of
	left%right
is if the '%' is the only recognized routing token left.

--dan dorough, systems programmer, ccdan@ucdavis.edu, (916)752-3420