[Nut-upsuser] Keeping the traffis on or off the list ?

[Charles Lepple]

On Feb 17, 2024, at 3:29 AM, Roger Price wrote:
> 
> Note that the Reply-To goes back to the original poster, not the list.
> Many mailing lists encourage the subscribers to "keep the list traffic on the list", rather than wandering off into private discussions.  The nut-upsuser setup has exactly the opposite effect.

Thanks for bringing this upstream change to our attention.

For context, NUT has been using Mailman through the Alioth mailing list services ever since the project used CVS and Subversion code hosting services from Alioth as well. While it is handy to not have to maintain all of the infrastructure of a mailing list server (especially in this age of complex requirements for successful mail delivery), it also means that we do not have a ton of visibility into the changes made upstream.

As Matus pointed out, the header-signing requirements are apparently forcing mailing lists to put in default values for some headers. While I am not sure I agree with the Mailman decision, this has made me aware of two settings that seem to have crept in when we weren't looking:

- first_strip_reply_to (currently: no)

- reply_to_goes_to_list (currently: poster)

Documentation: https://www.gnu.org/software/mailman/mailman-admin/node11.html

I have not yet adjusted reply_to_goes_to_list on the Mailman end, but I set the Reply-To header in my mail client in addition to putting the list in the To field. In theory, the first_strip_reply_to setting should allow my Reply-To header to pass through.

If that works, I am okay with changing reply_to_goes_to_list to actually point to the list. IMHO, if subscribers really want replies to their messages to include their address, they can go to the trouble of setting a Reply-To header that includes both the list and their personal address.

On Feb 17, 2024, at 5:09 AM, Roger Price wrote:
> 
> Debian doesn't do it.  Why should we?


Debian (and NetBSD, to Greg's point) are larger projects that can afford to make much more intentional choices about their infrastructure, such as mailing lists. They can also directly monitor the fallout from decisions such as not signing headers. If anyone is interested in taking on the project of maintaining a separate mailing list server for NUT, please get in touch with the -owner alias for this list. (If that sounds like an overreaction, but you would still like to help manage the current mailing lists, we could also use a few more moderators.)

[please use Reply-All as we get this all sorted out, thanks. No intentional asides to me; I will instinctively add back the list address as I have been doing manually for years when people email me directly about NUT issues.]

-- 
Charles Lepple
clepple at gmail