>>>>> "RSW" == Randall S Winchester <rsw@Glue.umd.edu> writes:
RSW> Say I have a list of 300 people who can post to a list, but only 30
RSW> people actually want to recieve mail from the list.
There are at least three options.
RSW> Is this easy to manage?
I guess that depends on how you define 'easy', but I think it's pretty easy
once you figure out how you're going to do it.
Here are your options:
Sign everyone up, but put the people who don't want to get anything in the
class 'nomail'. (set blah-list nomail email@example.com). This means the
people who can post are real subscribers; they just don't get any
messages. Set restrict_post appropriately:
Sign up the recipient users and put everyone else in a sublist:
auxadd blah-list posters firstname.lastname@example.org
then restrict posting to the main list and that sublist:
restrict-post is complicated because it tries to be backwards compatible.
The first element can be colon or tab separated, as in 1.94.4, but it's
really an array variable. Internally Mj2 tries to figure out what you were
trying to do when it comes to naming the restrict files, only it uses
sublists instead of files containing addresses. This means that you can
manage them remotely.
Sign up the recipient users, put everyone else in a sublist, make a file
explaining that only members can post, 'put' it with the name
'membersonly', and make an access rule:
!@MAIN && !@posters
This rule denies all postings from users who are not in the main list and
are not in the 'posters' sublist. They will (i.e. should, modulo bugs) be
sent the named denial message.
The latter is the most flexible; you can change 'deny' to 'consult' if you
want the list owners to be consulted. Also note that restrict_post isn't
backwards compatible when it comes to other requests ('who', 'get', etc.)
because I admit I really don't understand how it all works in 1.94.4 and
from what I do understand it doesn't really work in a sensible manner.
With access_rules you can describe exactly what you want.
And a final note: this is getting into less-tested territory. Here be
bugs. If you find any bizarre behavior, please send some debug logs my