Can I fix this groups.io problem with a send hook or something

[Cameron Simpson]

On 24Nov2020 12:09, Ian Zimmerman <itz at very.loosely.org> wrote:
>On 2020-11-24 17:22, Chris Green wrote:
>> > > I'm on a couple of mailing lists that have recently moved from
>> > > Yahoo Groups to groups.io.  For DMARC (so they claim) groups.io
>> > > change the senders E-mail address as follows:-
>> > >
>> > >     Sender Name <sender.name at btinternet.com>
>> > >
>> > > becomes:-
>> > >
>> > >     "Sender Name via groups.io" <sender.name=btinternet.com at groups.io>
>
>> It only happens to users from some domains. [...]
>
>It is because those senders subscribed with an address covered by a
>strict DMARC policy. If their address is not munged by the list, their
>messages will be bounced back to the list by almost everyone if they
>fail a DKIM check (and that's almost always). [...]
>
>There's no good automatic solution to build into a MUA for this, as far
>as I know. The best you can do is write a macro to extract the true
>sender, and feel lucky for using a MUA that allows that :-)

I have an unmangle step in my mail filer. I'll list the rules at the 
bottom of this message.

Basicly, for particular sources I hack the From: line around. For yahoo 
lists the original address was embedded in the From: header itself, for 
mercurial at mercurial-scm.org the Reply-To: header holds the original 
author, for users at lists.fedoraproject.org the X-Old-From: header holds 
it, for core-mentorship at python.org it is in the X-MailFrom: and so 
forth. So there's a suite of special purpose rules below.

The rule syntax below is "target condition...", so each of the:

    from:s/this/that/

lines is a "target" and the other lines constraint which messages it is 
applied to.

I imagine you could do something with procmail like this. See the 
"formail" command. If formail won't do it, a sed script probably will.

Note that this of course mangles the headers, a price I've been willing 
to pay for sanity.

Cheers,
Cameron Simpson <cs at cskk.id.au>

# undo Yahoo Groups DKIM/DMARC mangling
# if we ever check the validity of this stuff it will need to be before this step
from:s/"(?P<fullname>.*\S)\s+(?P<coreaddr>\S+@\S+)\s+\[(?P<listname>[^ ]*)\]".*<(?P=listname)(-noreply)?@yahoogroups.com>/"$fullname" <$coreaddr>/
                        sender:/@yahoogroups.com
                        from:/@yahoogroups.com>
from:s/"(?P<fullname>.*\S)\s+(?P<coreaddr>\S+@\S+)\s+\[(?P<listname>[^ ]*)\]".*<(?P=listname)@yahoogroups.com>/"$fullname" <$coreaddr>/
                        sender:/@yahoogroups.com
                        from:/@yahoogroups.com>
from:s/"(?P<coreaddr>\S+@\S+)\s+\[(?P<listname>[^ ]*)\]".*<(?P=listname)@yahoogroups.com>/<$coreaddr>/
                        sender:/@yahoogroups.com
                        from:/@yahoogroups.com>

from:s/.*/$reply_to/
                        sender:mercurial-bounces at mercurial-scm.org
                        from:/ via Mercurial <mercurial at mercurial-scm.org>

# undo Linux Australia header mangling
from:s/.*/$reply_to/
                        sender:linux-aus-bounces at lists.linux.org.au
                        from:linux-aus at lists.linux.org.au

from:s/.*/$x_mailfrom/  from:core-mentorship at python.org
                        x-mailfrom:/@
                        list-id:/core-mentorship.python.org/
from:s/.*/$reply_to/
                        from:(python-list at python.org|python-ideas at python.org|python-dev at python.org|tutor at python.org)

# fedoraproject.org x-mailfrom
from:s/ via users <users at lists.fedoraproject.org>/ <$x_mailfrom>/
                        from:(users at lists.fedoraproject.org)

# googlegroups x-original-from
from:s/ via users <users at lists.fedoraproject.org>/ <$x_original_from>/
                        from:vim_use at googlegroups.com