Re: [Emailcore] Ticket #5: G.5. Remove or deprecate the work-around from code 552 to 452?

[Jeremy Harris <[e-mail]>]

On 15/02/2021 23:46, Michael Peddemors wrote:
> Oh, to avoid confusion.. MagicMail SMTP returns..
> 
> rcpt_checks/max_rcpts.c:    return "452 Too many recipients (#4.5.3.1)";
> 
> However, the question remains.. why should it be temporary? We have always just done this according to RFC, but it doesn't seem logical, as the remote client/MTA will just queue up that batch for retry, no?

It depends what command is being responded to.

It it was a RCPT command, after several such had been accepted, a 4xx
should mean that the accepted batch are still ok for DATA, and the
remainders will need to go as a separate transaction.

Even as a response to end-of-data it's not totally unreasonable - I might imagine,
say, a disk-space check failing *because of* the number of accepted recipients.
Unusual, and I don't know if any MTA in existence does that, but possible.


On the other side of the coin, I can imaging anti-spam configured MTA deliberately
permanently-rejecting an entire message because it regards the attempted number
of RCPT commands as excessive.  So, both 4xx and 5xx are reasonable and
should mean what they mean for other "xx".
-- 
Cheers,
   Jeremy