Michael Häusler <[EMAIL PROTECTED]> said: >> I think you're describing the way it already works. If there's both a TEMP >> and a FAIL condition, worstcase returns TEMP. > > But I spoke of an OK and a FAIL. Worstcase will then return a FAIL > (which is reasonable I think). From dsn.c: > [...] > /* 2.. */ "Success", > /* 4.. */ "Persistent Transient Failure", > /* 5.. */ "Permanent Failure" > [...] > else if (has_2 && !has_4 && has_5) /* 2 and 5 */ > exitcode = DSN_CLASS_FAIL; > > But, you don't want to believe me, do you? ;-)
Oh, oh. Now that this sits before the insert_messages() call, you are correct that a good address will be dropped if there is a bad address. The former behavior was that the good address would receive mail, but the sender would be told of a failure on account of the bad address. Good catch. What's your suggestion for fixing it? Aaron
