>>>>> "EC" == Evan Champion <email@example.com> writes:
EC> auth 91b89f37 subscribe humour-list Shauna Lemons firstname.lastname@example.org
I don't understand this. That confirmation shouldn't have made it out
without bouncing. Hmmm. Perhaps sendmail took that as three addresses and
bounced the first two.
I think that previous behavior (before my big patch) would have done the
same thing, so I'll have to look elsewhere. (Though I'd appreciate people
running just alpha9 testing this, if possible.)
This is likely to be a problem (because of sendmail's possible treatment of
this as three addresses) and I don't see a simple way of fixing it. I'd
appreciate additional input from everyone else on this, as I'm beginning to
get lost deep within the mess of calls to valid_addr which don't really do
The problem here is that we don't have an RFC822 verifier, so we can't
simply check the legality of an address. We can strip it (with the
reasonably incorrect stripper that parse_addrs is) and check to see that
we've got something with an @ in it, but that still doesn't solve the
addr@host Full Name
which the stripper will leave unchanged! Ouch!
I don't think there's a solution to this without rewriting the address
parser. I'm just using Mailtools for the parsing in TLB, which tends to be
much, much more correct.
Now, I'll get busy making sure that the confirmation address gets stripped.