At 2:07 AM -0700 on 8/6/08, SM wrote about Re: retry question:
I expect MUAs to fail if there is any 4yz or 5yz response to a RCPT.
WHY? I have a message that is addressed to 10 people via the TO and CC Headers (I'll ignore the use of BCC). I attempt to send it and (if all 10 addresses were checked) get a mixture of accepted, 4xy, and 5xy responses. While I can see removing the 5xy'ed addresses from the headers, why should I be forced to destroy the 4xy'ed headers just to deliver the message to the others (and thus prevent the 4xy'ed ones from seeing any replies)? A much better way is to submit to the accepted addresses, warn the user of the 5xys so they can be MANUALLY removed, and then requeue the message for a later attempt to submit (with the already accepted during a prior attempt addresses not resent as RCPT-TOs).
