🚀 go-pugleaf

In general it seems much riskier to me to fail to warn people that they
are sending to an empty list. It seems like a great way to lose a lot of
mail. You must also remember that lots of postmaster mail is deleted without
ever being read, especially when it looks like a routine send failure of
some kind... Sending such traffic to the postmaster is therefore not an
effective solution in most cases.

It also seems to me that empty expansion lists are something that primary
list maintainers actually _do_ want to know about. True, it may be
inappropriate in some cases to return such information to the primary
list maintainer, but in other cases the primary list maintainer may need
this information to run the list effectively. The risk of the primary list
maintainer acting in an inappropriate fashion seem far less than the risks
of eliminating this information completely.

In any case, it is pretty easy to get PMDF to handle this in whatever
way you like. You can arrange for any alias to override the error return
addresses in a message. Just set things up so that the externally visible
list address sets the error return addresses to either the local
postmaster or <>. This will require another indirect file but that's no
big deal. And the bogus address on the list will eliminate the possibility
of the list ever emptying completely. The bitbucket channel can be used to
process such bogus addresses.

				Ned