Re: Envelope tests on null or illegal MAIL FROM (non)addresses

[Ned Freed <[e-mail]>]

> On Wed, 2006-06-14 at 08:03 -0700, Ned Freed wrote:
> > FWIW, I'd be in favor of having a null MAIL FORM result in an empty string
> > being tested regardless of which ADDRESS-PART is specified.

> agreed.

> >  As for invalid
> > MAIL FROM values, I'm inclined to say that :all should test against the
> > literal string inside the <>s and :domain and :localpart should test
> > against the empty string. This way you can test for specific bogons and
> > handle them accordingly.

> as a user, I would expect it to do the simple thing, and split on the
> (rightmost) @.  but by requiring syntax correctness, we protect users
> from themselves, ie. if they write a match expression which didn't
> consider the possibility of some particular external garbage.

> wouldn't it be better to say that :localpart and :domain never match
> when the address is a bogon, though?  I don't see what we gain by making
> it "alias" into an empty sender, even if the recommended and natural way
> of checking for the empty sender is with :all.

Good point. I agree that domain and localpart matches should never match when
dealing with an unparseable address.

				Ned