[postgis-devel] DMARC/DKIM mitigation on maling lists

[Magnus Hagander]

It is not, no. We use pglister (homegrown, though has a couple of
other users by now) -- https://gitlab.com/pglister/pglister

I think the test was more to indeed pinpoint that the problem is in
mailman, and not in some other part :)

//Magnus

On Fri, Oct 27, 2023 at 4:25 PM Laurențiu Nicola <lnicola at dend.ro> wrote:
>
> Interesting, my message to Magnus' test list arrived with my valid DKIM signature and no extra Sender header.
>
> But this is not Mailman, is it?
>
> Laurentiu
>
> On Fri, Oct 27, 2023, at 16:06, Laurențiu Nicola wrote:
> > Hello,
> >
> > I don't know much about the meaning of email headers, but I noticed
> > that osgeo.org adds itself as the Sender, which might explain the DKIM
> > signature mismatch I've noticed.
> >
> > The whole Mailman and DKIM situation seems pretty complicated.
> > According to https://wiki.list.org/DEV/DKIM, which is old but discusses
> > the viable alternatives, the preferred method seems to be simply
> > stripping the original DKIM signature. This is equivalent to keeping
> > the original one, as a broken and a missing signature are supposed to
> > be handled exactly the same. Re-signing the Mailman messages might also
> > be an option.
> >
> > I don't have a strong opinion, I just mildly dislike the rewriting of
> > From because it messes up my address book.
> >
> > If we can disable the Sender rewriting, that might be an interesting
> > exercise, otherwise, I suspect we might as well ignore DKIM.
> >
> > Laurentiu