--On Saturday, April 16, 2016 04:40 +1000 Mark Andrews
<marka(_at_)isc(_dot_)org> wrote:
verify != canonicalization. One checks existance the other
doesn't.
If SMTP servers provided a canonicalization service
Case insensitive
CANON loCal(_at_)eXamPle(_dot_)CoM
200 local(_at_)example(_dot_)com
Google's remove dots + case insensitive would result in
CANON lo(_dot_)C(_dot_)al(_at_)eXamPle(_dot_)CoM
200 local(_at_)example(_dot_)com
Case insensitive base w/ plus suffix
CANON loCal+IETF(_at_)eXamPle(_dot_)CoM
200 local+IETF(_at_)example(_dot_)com
independent of whether local(_at_)example(_dot_)com exists or not
But, as an email provider, I might refuse to tell you about how
I would interpret addresses that don't exist because I might
have different rules for different mailboxes (especially when
non-ASCII local parts are concerned). On the other hand, if I
refused to tell you about _any_ addresses that didn't exist,
we've have the VRFY problem all over again.
In addition, as Ned and others have pointed out,
CANON local+IETF(_at_)example(_dot_)com
Might be just
local(_at_)example(_dot_)com
(or not, depending on one's actual purpose and what is going on
in the server and maybe the client -- whether those two are
equivalent might depend on the day of the week.) And there are
reasons to think that
EQU loCal+IETF(_at_)eXamPle(_dot_)CoM local+ietf(_at_)example(_dot_)com
might be a better idea than your hypothetical CANON command.
I'm _not_ making a proposal, only trying to suggest that anyone
who thinks this is an easy and straightforward situation, even
if the question is posed to the delivery SMTP server rather
than, e.g., coded as rules in the DNS, just hasn't thought
deeply enough about the issues (or doesn't understand them).
john
_______________________________________________
ietf-smtp mailing list
ietf-smtp(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf-smtp