Autor: Sander Smeenk Data: A: exim-users Assumpte: Re: [exim] OT(?): SRS vs. SORBS sending mails with '<>' sender
Quoting Heiko Schlittermann (hs@???):
> what I understand from SRS is:
> If I use SRS for outgoing messages, bounces to messages I sent should be
> addressed to SRS-"encoded" recipients, not to real real recipients.
In my honest opinion, bounces should always go to the envelope sender
address. E.g. the one specified at SMTP time, the one you've verified to
have MX-records using verify = sender.
A DSN originates from a null sender to stop loops, normal email should
(again in my honest opinion) never be sent from a null sender address.
So i drop every DSN to addresses that are 'protected' by SRS and that
doesn't meet my SRS policy.
> Who is wrong here? Should SORBS use valid non-null sender addresses or
> should I forget thinking about SRS?
I'd say SORBS.
> [ In case SORBS is wrong - does anybody here know some folks from SORBS
> to ask for a fix? ]
Better would be to implement a whitelist in your SRS-check, like i did.
You'll find it hard to convince people to change their setup because you
have a problem.
I can whielist certain hosts so they pass the test. In my case it was
necessary because of broken MTA's doing callouts using null sender
address where they should've used postmaster@???. F.e.
the mysql mailinglistserver does that.
