Re: question about sender_dependent_default_transport_maps syntax
Wietse Venema porcupine.org> writes: > > Fabio Sangiovanni: > > Hi all, > > > > from the docs of sender_dependent_default_transport_maps: > > "Note: this overrides default_transport, not transport_maps, and > > therefore the expected syntax is that of default_transport, not the > > syntax of transport_maps. Specifically, this does not support the > > transport_maps syntax for null transport, null nexthop, or null email > > addresses." > > > > And from the docs of default_transport: > > "The :nexthop destination is optional; its syntax is documented in the > > manual page of the corresponding delivery agent." > > > > Is someone willing to clarify this a little? My question raises from the > > fact that I used this configuration, and it worked: > > > > main.cf: > > sender_dependent_default_transport_maps = > > hash:/etc/postfix/sdd_transport_maps > > > > # not-null transport, null nexthop > > /etc/postfix/sdd_transport_maps: > > mysender domain.tldmytransport: > > You need to specify a nexthop destination only if you want to > override the default. The default nexthop destination is the domain > portion of the recipient address. > I'm perfectly aware of this. To leave the nexthop destination at default I need a null nexthop. This is against what the documentation says about sender_dependent_default_transport_maps ("this does not support the transport_maps syntax for null transport, null nexthop, or null email addresses."). What I'm pointing out here is a possible typo in the docs. > > /etc/postfix/master.cf: > > mytransport unix - - n - - smtp > > -o smtp_generic_maps=hash:/etc/postfix/generic > > > > What am I missing here? > > It works but does not to match yesterday's requirements: > > I need to rewrite the address of some particular recipients, > but just for messages with envelope sender different from the > null sender. In other terms, I need to redirect messages for > some recipients in a list (towards other, remote, addresses), > but only if the envelope sender is not <>. In other terms, I > need to redirect messages for some recipients in a list (towards > other, remote, addresses), but only if the envelope sender is > not <>. If the envelope sender is <>, I need messages to follow > the standard route towards the internal host, even for the > recipients on the list. > > Before I proceed to write down a configuration that you don't need, > perhaps you can explain the *problem* that you are trying to solve, > instead of your solution to route <> envelopes differently. If I wanted further advice about yesterday's topic I would have answered whithout changing the subject, as I did this morning. But in this case, I don't. I just asked for clarifications about the docs. And that's exactly the reason I started a new thread. As far as I'm concerned, our agreement still stands, and I'd appreciate if you could help with configuration about that matter, since *those* requirements, as reported yesterday, haven't changed. Fabio
Re: question about sender_dependent_default_transport_maps syntax
Fabio Sangiovanni: > Hi all, > > from the docs of sender_dependent_default_transport_maps: > "Note: this overrides default_transport, not transport_maps, and > therefore the expected syntax is that of default_transport, not the > syntax of transport_maps. Specifically, this does not support the > transport_maps syntax for null transport, null nexthop, or null email > addresses." > > And from the docs of default_transport: > "The :nexthop destination is optional; its syntax is documented in the > manual page of the corresponding delivery agent." > > Is someone willing to clarify this a little? My question raises from the > fact that I used this configuration, and it worked: > > main.cf: > sender_dependent_default_transport_maps = > hash:/etc/postfix/sdd_transport_maps > > # not-null transport, null nexthop > /etc/postfix/sdd_transport_maps: > mysen...@domain.tldmytransport: You need to specify a nexthop destination only if you want to override the default. The default nexthop destination is the domain portion of the recipient address. > /etc/postfix/master.cf: > mytransport unix - - n - - smtp > -o smtp_generic_maps=hash:/etc/postfix/generic > > What am I missing here? It works but does not to match yesterday's requirements: I need to rewrite the address of some particular recipients, but just for messages with envelope sender different from the null sender. In other terms, I need to redirect messages for some recipients in a list (towards other, remote, addresses), but only if the envelope sender is not <>. In other terms, I need to redirect messages for some recipients in a list (towards other, remote, addresses), but only if the envelope sender is not <>. If the envelope sender is <>, I need messages to follow the standard route towards the internal host, even for the recipients on the list. Before I proceed to write down a configuration that you don't need, perhaps you can explain the *problem* that you are trying to solve, instead of your solution to route <> envelopes differently. Wietse
question about sender_dependent_default_transport_maps syntax
Hi all, from the docs of sender_dependent_default_transport_maps: "Note: this overrides default_transport, not transport_maps, and therefore the expected syntax is that of default_transport, not the syntax of transport_maps. Specifically, this does not support the transport_maps syntax for null transport, null nexthop, or null email addresses." And from the docs of default_transport: "The :nexthop destination is optional; its syntax is documented in the manual page of the corresponding delivery agent." Is someone willing to clarify this a little? My question raises from the fact that I used this configuration, and it worked: main.cf: sender_dependent_default_transport_maps = hash:/etc/postfix/sdd_transport_maps # not-null transport, null nexthop /etc/postfix/sdd_transport_maps: mysen...@domain.tldmytransport: /etc/postfix/master.cf: mytransport unix - - n - - smtp -o smtp_generic_maps=hash:/etc/postfix/generic What am I missing here? Thanks a lot Fabio Sangiovanni