Re: 2821bis AUTH48 fix (?)

Tony Finch wrote:

Hi Tony,

> > I'd guess it was something in the direction of sending DATA
> > after all RCPT TO were 4xx or 5xx rejected (= empty list of
> > 2xx RCPT).
 
> The spec is fairly clear about what to do in that situation.
Yes, that's the base of my guess, John wrote "outside the spec."
Using 4xx and 5xx at all after RCPT TO or DATA is not "outside".

 Frank