Errors-to: owner-drums@CS.UTK.EDU
From: <Harald.T.Alvestrand@uninett.no>
Message-Id: <199506040707.DAA01065@CS.UTK.EDU>
To: John Gardiner Myers <jgm+@CMU.EDU>
cc: drums@CS.UTK.EDU
Subject: Re: RFC 821 problem list
In-reply-to: Your message of "Wed, 31 May 1995 13:46:09 EDT." <8jn_hVu00WBwM0wcN6@andrew.cmu.edu>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Sun, 04 Jun 1995 08:58:15 +0200
Sender: Harald.T.Alvestrand@uninett.no

John,
I think you are thinking about page 43 of RFC 821, which says:
....
            recipients buffer

               The maximum total number of recipients that must be
               buffered is 100 recipients.
....
         Errors due to exceeding these limits may be reported by using
         the reply codes, for example:
....
            552 Too many recipients.


In this case, an implementation that follows 1123 and examines only the first
digit of the reply code will bounce the message.

If the response started with 4, the same mailer would retry exactly the
same transaction; since all the too-many recipients are intended for delivery
to this exact system, the number will not go down on a retry, and the error
is likely to remain permanent.

An "intelligent" mailer could take the recipient list apart and try it
again in 100-recipient chunks, but this would require the parsing of more
than one digit of reply code.

My conclusion:

- "552 Too many recipients" belongs on the problem list (that I sure hope
  the chair or someone is busily typing up!)

- It is a permanent error to anyone who doesn't understand the exact code,
  so a 5xx code is appropriate.

             Harald A