🚀 go-pugleaf

Paul,

neither rfc822 nor rfc976 specify that % does what the %-hack do.  In fact,
both specify other methods to do routing.  That means, both specs expect
the local-part to only be paid attention to by the 'local' machine, not
by machines along the route.

In:

	user%a@some.dom.ain
and	some!long!path!with!a%percent

the local part *shouldn't* be evaluated until it reaches the destination.
In the '@' case, until it reaches some.dom.ain, and the '!' case until
it reaches 'with'.  Whatever either of those places wishes to do with
a local-part like "a%b" is up to them BECAUSE IT IS NOT SPECIFIED BY
EITHER SPEC!

Now, I didn't read your posting very closely and it's possible that you
said exactly that in your posting.  But I think I caught a few statements
to the otherwise.

Oh, I suppose the next question is what to do with 

	some!long!path!a%b@some.dom.ain

RFC976 is the applicable spec.  (according to rfc822, the "some!long!path!a%b"
is *all* local part and get's evaluated according to some.dom.ain's rules
for evaluating such things).  The rfc recommends first that mailers not
use mixed adresses internally, instead preferring something like

	some.dom.ain!some!long!path!a%b
or	some!long!path!some.dom.ain!a%b

but if you must, to treat such an address as in rfc822 -- the local part
is the local part and isn't evaluated until it reaches some.dom.ain
(i.e. the "some.dom.ain!some!long!path!a%b" interpretation -- effectively)

It's right there in the specs in black&white (or purple&brown with the
right toner/paper combination ...)
-- 
<---- David Herron -- The E-Mail guy                         <david@ms.uky.edu>
<---- ska: David le casse\*'      {rutgers,uunet}!ukma!david, david@UKMA.BITNET
<----                   A misplaced Kansan trapped in the heart of Kentucky,
<---- the state where it is now illegal to water your lawn on the wrong day.