Thanks for the idea. I have created a dump using tcpdump and analyzed this using wireshark.
I can see the following transmission: *>> get i...@foodtruckbestellen.be <i...@foodtruckbestellen.be>.* *<< 200 SRS0=sgig=PE=domain.com <http://domain.com>=i...@otherdomain.com <i...@otherdomain.com>.* *>> get srs0=sgig=pe=domain.com <http://domain.com>=i...@otherdomain.com <i...@otherdomain.com>.* *<< 200 srs0=sgig=pe=domain.com <http://domain.com>=i...@otherdomain.com <i...@otherdomain.com>.* So apparently, Postfix receives the SRS rewritten address, but then it requests another SRS rewrite from PostSRSd using the already rewritten address (lowercased). Any idea how this could be possible and how it could be fixed? Thanks for the feedback so far. It looks like we're on the right track to finding the problem. Kind regards Wesley Op do 27 dec. 2018 om 21:28 schreef Wietse Venema <wie...@porcupine.org>: > Wesley van Synio: > > I still can't find the cause of the problem. > > > > When I disable the sender_canonical_maps then SRS is not working, > > and when I then use i...@domain.com, then case folding is not happening > for > > Return-Path. > > So the problem must be in the SRS rewriting or in the connection between > > PostSRSd and Postfix? > > > > But I already asked the PostSRSd author and also checked the source code > > and it should work... > > Also, the log output mentions a correct rewrite without case folding. > > > > How can I easily debug what Postfix sends and/or receives to/from > PostSRSd > > so I can debug further? > > I'm thinking the problem must be located somewhere around there...? > > > > Any other insights would be helpful! > > Are you using the TCP map for sender_canonical envelope mapping? > In that case, can you log the input and output? Use a network > sniffer, for example, > > tcpdump -i lo -s 0 -w /tmp/foo port 12345 > > Run it for some time, then stop with ctrl-c. > > View with "strings /tmp/foo" or more sophisticated tools. > > Wietse > -- *Synio* *Ilgatlaan 9* *3500 Hasselt* *0477 71 79 25*
