>>>>> "DW" == Dave Wolfe <email@example.com> writes:
DW> This seems backwards.
It's a toss up. The actual core function takes two addresses, but since
it's impossible to determine where one address ends and another begins, I
can only accept one on the command line. The parser extracts the other
from the header and passes them to the core. Which one goes where was the
result of a coin toss.
This problem comes up again with the "rezubscribe" function, which changes
a zubscribed address. (Which I also haven't written yet...)
DW> At least one case where this would be useful is to subscribe a "receive
DW> only" address (a forwarding mailbox that remains constant while the
DW> actual mail reading/posting site changes) and alias the variable
DW> posting addresses.
That makes sense to me.
Now that I think about it, with the current implementation it is actually
possible to just figure it out without having the user specify. Since one
of the addresses must be subscribed and the other must not be subscribed,
it's pretty easy to figure out which goes where.
DW> Or does the Reply-To hole still exist in mj2?
Hole? You don't mean "security hole", do you? Yes, Majordomo pays
attention to Reply-To: just as it always has.
DW> I realize there's a security issue involved, but it seems to make the
DW> feature useless for a large section of the problem set.
Any security issues can be resolved by confirmation to the subscribed
address. This is in general the _only_ way we can resolve any security