Hello Stephan,
First of all, thank you for dovecot-sieve implementation.
I see that you are adding X-Sieve header to all outgoing e-mail in
cmd-redirect.c
/* Prepend sieve version header (should not affect signatures) */
rfc2822_header_field_write(f, "X-Sieve", SIEVE_IMPLEMENTATION);
How do you look at adding one more header, for example X-Sieve-Forward-From:
Now, when redirect action is in use - mail forwards using sendmail command line
interface. The problem that there is no suitable way to know which user's
script has generate redirect. Received: header may contain envelope receiver but
this is valid if only 1 "rcpt to:" command was specified on SMTP
session.
This invokes problem when forwarding mail to non-local domains, breaking such
things like SPF. I'd like to setup SRS for fixing this, but I don't know
sender("sieve local user") mail. Adding one header with valid local
user e-mail address will fix this.
Google add 2 headers for this -  X-Forwarded-To: and X-Forwarded-For: .
What do you think about this ?
Nikita Koshikov schreef:> Hello Stephan, > > First of all, thank you for dovecot-sieve implementation. > > I see that you are adding X-Sieve header to all outgoing e-mail in cmd-redirect.c > > /* Prepend sieve version header (should not affect signatures) */ > rfc2822_header_field_write(f, "X-Sieve", SIEVE_IMPLEMENTATION); > > How do you look at adding one more header, for example X-Sieve-Forward-From: > > Now, when redirect action is in use - mail forwards using sendmail command line interface. The problem that there is no suitable way to know which user's script has generate redirect. Received: header may contain envelope receiver but this is valid if only 1 "rcpt to:" command was specified on SMTP session. > This invokes problem when forwarding mail to non-local domains, breaking such things like SPF. I'd like to setup SRS for fixing this, but I don't know sender("sieve local user") mail. Adding one header with valid local user e-mail address will fix this. > Google add 2 headers for this - X-Forwarded-To: and X-Forwarded-For: . > > What do you think about this ?I have no objection, but then again my experience on this subject is pretty bleak. Does anyone else have comments? Regards, -- Stephan Bosch stephan at rename-it.nl
Maybe Matching Threads
- Unexpected behavior with sieve_vacation_to_header_ignore_envelope = yes
- outlook and redirect sieve
- openssh-unix-dev DMARC-related settings (was Re: scattered thoughts on connection sharing)
- sender rewriting scheme
- i386 version of Perl getting installed on x64 system?