[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: nonmember=reject settings

From: Karl Berry
Subject: Re: nonmember=reject settings
Date: Tue, 16 Feb 2021 16:57:55 -0700

    > that we could keep updated with education and reasons for what
    > backscatter spam is and the problems it causes.  

I wrote a terse explanation long ago on our "recommended mailman
configuration" page:
(bottom of page)

    The question I have is, which lists are ok to leave as
    nonmember=reject, if any? 

IMHO we should never have nonmember=reject because of backscatter.
But sometimes we have given in. Here's a preliminary list of
ones that do (the "2" is the numeric value for Reject):

school-free-software-private true  (nonmember special 2)    
paparazzi-devel false  (nonmember special 2)    
reproduce-devel false  (nonmember akhlaghi 2)    
info-gnuastro false  (nonmember akhlaghi 2)    
gnuastro-devel true  (nonmember akhlaghi 2)    
libassh-users true  (nonmember 2)    
igraph-help true  (nonmember 2)    
foliot-user true  (nonmember 2)    
bug-g-golf true  (nonmember 2)    
bug-foliot true  (nonmember 2)    

And here is more information about these lists:

The "special" means we had actually communicated with the list owner at
some point and they really didn't want to change. I think the school-*
list is FSF owned?

The "akhlaghi" lists are those owned by (a super
nice guy and big fsf supporter, as well as an active gnu maintainer).
He also did not want to change.

The others are the "new" ones (some of them might not be so new) that I
wrote about that started this thread.

BTW, the true/false is whether the list uses listhelper.

    How about nonmember=discard?

OTOH, there are a number of lists with nonmember=discard, for lists that
are supposedly private and "invitation only". It seems more or less a
legitimate case to me, at least not enough to spend time "explaining" to
list owners. I don't have a list of those at hand, but I wouldn't try to
change that now. I suggest focusing on reject first, since that's the
worst offender.

There is one other super-special case:
dtas-all true  (nonmember 0)    

0=accept! Yes, this list owner was absolutely fanatical about having
nonmember=accept, so that anything at all, including spam, would come
through with *no* delay. I've looked in on the archives from time to
time, and it seems only one or two spams have arrived. So ... shrug.

BTW, in case it's useful at any point, I have a monthly cron job that
dumps all the mailman configurations to /home/list/tmp/dumpallconf (huge
directory) and then does some sloppy text-based parsing to find stuff
like the above. I built up more scripts around it (all under ~list), but
I won't bother you with them.  I imagine your scripts are much cleaner :).

All the best,

reply via email to

[Prev in Thread] Current Thread [Next in Thread]