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

[Hector Santos <[e-mail]>]

On 3/28/2021 10:00 PM, John C Klensin wrote:
> Hector, RFC 5321 said, in 2008, to use 452 for this case. The only 
> justification I can see for leaving in the text that Alexey and 
> others have proposed removing would be if leaving it there a bit 
> longer would cause you to bring wcSMTP into compliance. If the 
> reality is that either you know about the change and will get around 
> to it someday regardless of what 5321bis says or that nothing we say 
> (or don't say) in 5321bis will make any difference, then it seems to 
> me that we should eliminate confusing text that no longer applies to 
> a significant number of other MTAs. Which is it? Are you suggesting 
> retaining that text and, if so, why?

I agree it would be better to use 452  and keep the expected SMTP 
state machine working with the 4yz (temporary) and 5yz (permanent) 
logic.  I have no comments regarding the suggested text changes.

I probably thought about it back in 2821bis and decided to leave it 
alone since it was an 821 design. Making it 452 might have created an 
unknown situation at the time.

On the receiver side, for the next update, I will make a change to use 
452 with an undocumented (why muddy the h2o) switch option to use 552 
just in case a revert is necessary.

On the client side, the SMTP client will permanently fail the 
transaction when a 5yz is seen for RCPT. No special case for 552. For 
our list server when prepared to go direct to SMTP, a 55z code will 
make the list member inactive. With 452, an error count is incremented 
for the member but not made inactive.  I need to relax this. The list 
server should check for 552 and not make the user inactive.  As long 
as no hash treatment to the limited list members is done, I don't see 
any other impact.

Thanks


-- 
Hector Santos,
https://secure.santronics.com
https://twitter.com/hectorsantos