On Thu, Apr 9, 2009 at 11:21 PM, Victor Duchovni <victor.ducho...@morganstanley.com> wrote: > On Thu, Apr 09, 2009 at 11:09:17PM +0200, Daniel Hahler wrote: > >> > propagate_unmatched_extensions = canonical, virtual
I was referring to "propagate_unmatched_extensions = canonical" here. >> But that would also include the whole of virtual delivery, no? (e.g. >> to virtual_domains etc > > Yes, of course. When some downstream domains can't handle recipient > extensions, you need to drop the extension for all downstream domains. And I've meant my own virtual_domains handling here: it does not get passed to downstream, but stays on the server (and the extension should not get dropped). >> Don't you think that smtp_generic_maps is a better approach? > > With smtp_generic_maps, you could selectively drop the extension, > provided you have a list of addresses to which the extensions may have > been propagated by upstream virtual rewriting. In this case you would > use identity smtp_generic_maps entries for the unextended addresses to > suppress the extension. That's what I'm doing now - currently for any gmx.* (extended) recipients. Thanks, Daniel -- http://daniel.hahler.de/
