Hi, > Perhaps with smtp_generic_maps, but not with canonical > maps
Which raises something I've wondered before: why is it that the generic_maps cannot be separately setup for header/envelope, and for sender/recipient, like can be done with canonical_maps? I've been struggling with this in the past, and ended up renaming everything upon entry and casting it back through the virtual table, but it felt like generic_maps would have been a better way, had it distinguished the things it would substitute. -Rick
