Re: [Nmh-workers] Garbage collect SENDMAILBUG code

[valdis . kletnieks]

On Wed, 31 May 2017 22:21:35 -0400, Ken Hornstein said:
> There is code in mts/smtp/smtp.c with the following comment:
>
> #ifdef SENDMAILBUG
> /*
>  * It appears that some versions of Sendmail will return Code 451
>  * when they don't really want to indicate a failure.
>  * "Code 451 almost always means sendmail has deferred; we don't
>  * really want bomb out at this point since sendmail will rectify
>  * things later."  So, if you define SENDMAILBUG, Code 451 is
>  * considered the same as Code 250.  Yuck!
>  */
> #endif
>
> I think this is long-obsolete, and I propose we remove it.  Objections?

Not just long-obsolete. That code has been incorrect since it was written.  The
451 being handed to *you* during an SMTP conversation means that you tried to
hand off an email to Sendmail, and it had hit "out of processes" or other temp
fail.

In other words, if *you* get a 451, Sendmail has *not* queued it up and won't
rectify things later - it's *your* job to queue and retry later.  Something that
nmh has traditionally sucked at.

(On the other hand, if sendmail encounters a 451 from the *next* system in
the path, and sends you a Delivery Status Notification saying so, *that* case
can be ignored as Sendmail will retry).

pgpJrVLNv3isC.pgp
Description: PGP signature