At 8:10 pm +0100 2004/08/14, Philip Hazel wrote: >On Sat, 14 Aug 2004, Giuliano Gavazzi wrote:
>
>>sorry if what I'll say is naive, but surely if we only cache callouts
>>overquota errors should give no problems, as quota should give an
>>error only at the DATA phase and this phase is never reached for
>>callouts.
>
>No, what the OP wants is to reject at RCPT time if the mailbox is over quota.
>
ah, I missed that. But callouts could be detected, even if not
distinguished from bounces at RCPT time, and the reject could be
delayed to DATA in this case.
I tend not to mess with possible callouts by subjecting them to just
the customary recipient verification. Now, the (Christiaan, I
imagine) OP's case is not known to me, so this might result
unsuitable.
