Odhiambo Washington writes:

Thanks, Florian, but why should I go that way when there is a link to sign in with google, which I already have?

Because the mailman3.org system is in beta (to be generous). Google sign-in is "supposed" to work, but doesn't. Try going through Persona with your Google account, as Simon suggested. Worked for me.

Are there people who have installed MM3 on FreeBSD? And integrated with Exim?

I think Turnbull has his running with Exim, but on Linux.

I got it running but that was a long time ago, pre-release (and pre-persona), and for a limited-time low-traffic list. Since then I've not had time to upgrade, and Postorius/HyperKitty weren't tested at all except for mass subscribe IIRC (although I might even have done that by mailman.client, as it was called then).

I see no reason why the Exim integration would change: https://gitlab.com/mailman/mailman/blob/master/src/mailman/docs/MTA.rst#exim As far as I can recall the MM3_UID and MM3_GID are no-ops, as Mailman 3 communicates entirely by sockets. Exim has no need to know. (The variables are a holdover from Mailman 2 where pipes are used, so Exim needs to run Mailman with the appropriate privileges IIRC.)

I'd love to have someone share their notes on installation on FreeBSD.

Mailman core really shouldn't be OS-specific (except to the extent that a supported MTA and Python are available). All communication takes place via (Unix-domain) sockets.

I'm not sure that things won't be a little hairier for HyperKitty and Postorius, but again, as long as Python runs, they should work the same on any system that supports sockets (that's one of the advantages of the current architecture -- zeroconf, not yet, but configuration is getting easier).

The only thing I can think of that might be vendor-specific is the location of Mailman itself, but at this point in time I see no reason to install it anywhere but /usr/local.

The real problems that people have are plain ol' bugs.