2020.12.09 20:44 "Re: [Tiff] Change of mailing list DMARC related setting", by Greg Troxel

(Yes, I understand DMARC.)

how lucky you are :-) I'm afraid I cannot turn your below suggestions in practice however. I can't see this dmarc_mitigate_unconditionally option in the mailman UI, at least not clsoe to the option I modified. Perhaps this was added in a later version than the one we have (2.1.34), or it is somewhere else in the UI.

I found out about it from reading the doc page you linked to. I have seen mailman tend to rewrite from only for domains that have dmarc policies, generally.

I let you add your suggestions in https://trac.osgeo.org/osgeo/ticket/2475 where you could find more qualified people than me for that subject. I believe the system administration committee would also welcome competent people on the subject.

ok, will comment there.

What is strange is that the libtiff mailing list seems indeed to apply dmarc mitigation unconditionanlly whereas the GDAL mailing list where I turned "munge from" to only does this for some people. So they are probably configured differently.

I suspect it is the setging I mentioned.

We're trying to avoid people from being "censured" as they are currently with their email being marked as spam by popular email providers.

Yes, I understand that.

The way DKIM works is that senders apply a signature and receivers can then say "was this email signed by the originating domain".

DMARC adds a policy "if it's From: this domain and DKIM validation fails, reject it".

Mailman, when configured to add things like "[tiff]" to the subject line, and things at the end, modifies the message in such a way that the dkim signature no longer validates. (gnupg signatures are also invalidated.)

So if you turn off subject tags, and turn off footers, then there is no need to work around dmarc because the message when it arrives will pass validation. Neither of these things is necessary because the information about which list is in the List-Id header and unsubscribe etc. is also in headers.