> will get through. And the worst case is the same as if we just didn't accept
The worst case is that some MTAs might simply not check whether the parts of
the address will fit into the buffer space which is allocated for them, and
because nobody in their sound mind would want to use addresses which exceed the
limits, such problems may well have gone unnoticed for a long time, until some
hacker discovers them as a method for making denial-of-service attacks, or
possibly even worse.
> What brought on this bout of anal-retentiveness, anyway?
I don't like the limitations in RFC821 any more than you do, but still I
believe that until that standard is officially revised, anyone who implements
the SMTP protocol should obey the must's and must not's of that RFC to the
letter, at least in the default configuration of the software.
Anyway, now that Jason has solved most of these issues in a satisfactory way,
it is certainly not wise to waste much more time with discussing these
"corner-case RFC grey-area compliance" matters, as Jason calls them; we will
document those problems which don't have satisfactory solutions yet for the
BUGS file and then move on to more important matters.