Re: [Evolution] Proposal: to eliminate the digest option for the Evolution mailing list

[Patrick O'Callaghan <poc usb ve>]

On Thu, 2017-03-16 at 01:22 +0100, Ángel wrote:
> On 2017-03-14 at 09:54 +0000, Pete Biggs wrote:
> > If digests are kept, are other mitigations possible. Like:
> >
> >     Is there a way to enforce mime digests?
> >
> >     And is there a way to make the boilerplate at the top of the
> > digest more "robust" on the issue of replying to a digest?
> >
> >     Is it possible to use References: headers to send all digest
> > replies to moderation?
>
> These are some digest Message-IDs:
> mailman 18 1489579205 7551 evolution-list gnome org
> mailman 19 1489579205 7551 evolution-list gnome org
>
> a search for
>  References: *mailman.*.evolution-list gnome org
> should probably do it.
>
> Given that the email received should already have a Message-ID, I expect
> few false positives. It is possible that in some cases the incoming
> email didn't include a Message-ID and was added by gnome MTA, but those
> shouldn't contain the mailman substring.
>
> The only false positives I'm coming up with is for someone purposefully
> an mailmain administrativia into the mailing list (eg. he is quoting the
> list Welcome message to point out a typo)

Even if we assume this is a reliable way to catch replies to digests,
there is nothing in the Mailman admin interface to specify filtering
criteria, other than a) non-membership of the list, or b) message body
is too large (40k currently). Such messages are held for moderation, so
filtering, including spam filtering, is being done manually by your
trusty moderators. It averages 2 or 3 per day so not a huge burden at
the moment.

> >     Is it technically possible to remove the ability to turn on
> > digests, but allow current users to keep receiving them?
>
> Hiding the UI at the subscription step may be enough. You could hide it
> at the user preferences too if you want to be more stubborn about it.
> This would require editing server files, so no, I don't think there's a
> mailman preference that can be readily used per list.

There isn't.

poc