
On Wed, Mar 22, 2017 at 2:11 AM, Niall Douglas via Boost
Still, I've never had this problem with the previous infrastructure. Maybe the email clients tend to not send the second email to the From address when there is Reply-To. For example, the std-discussion and std-proposals mailing lists use this scheme:
From: real original sender address To: mailing list address (e.g. "std-discussion@isocpp.org"
) Reply-To: mailing list address (e.g. std-discussion@isocpp.org) No Cc headers.
Can we use this scheme?
That was the previous scheme which caused the DMARC failure because DMARC won't allow this ML to send email From: an address it does not own.
Well, I'm not sure I understand the details, but I can see the problems with DMARC were resolved for isocpp.org lists at some point: https://groups.google.com/forum/#!topic/boost-steering/EcKn2yA9ip4 And I can see that now the emails from those lists contain the original sender address in the From header. I'm not sure how exactly that was achieved, but I would like this list behave the same way.
Regarding your problems, I'm on Thunderbird and I'm finding no problem at all with the new ML config. Somebody obviously told Thunderbird about the new munging option for Mailman and it all "just works".
The real problem seems to me to be gmail having stupid handling.
As I understand, this is not a client problem per se, since it is supposed to add CC to a reply when the user hits Reply All. Just as the mail server is supposed to deliver the message to all recipients. It so happens that mail delivery between gmail accounts sometimes happens faster than the roundtrip between gmail and the ML server, which makes the email from the ML a duplicate and lost. You're not seeing the problem probably because there are no other participants from nedprod.com, so the problem is less likely to appear for you.