Werner LEMBERG wrote: > > In the meantime I will change the config to this one. I think that > > will prevent the sending of the branch change notification but will > > still preserve the individual commit notifications. > > > > mailinglist = > > commitList = groff-com...@gnu.org > > This works, thanks. However, there are some issues. After a commit, > mails are sent twice to `groff-commit': > > Subject: [Groff-commit] [groff] 01/01: * tmac/groff_mdoc.man: \ > Improve the manual page template. > From: Werner LEMBERG <w...@gnu.org> > To: groff-com...@gnu.org > Date: Sun, 16 Feb 2014 07:28:09 +0000 > Sender: groff-commit-bounces+wl=gnu....@gnu.org > Reply-To: Ingo Schwarze <schwa...@usta.de>
Note: Sender: groff-commit-bounces+wl=gnu....@gnu.org I think that is the second extra one. > Subject: [Groff-commit] [groff] 01/01: * tmac/groff_mdoc.man: \ > Improve the manual page template. > From: Werner LEMBERG <w...@gnu.org> > To: groff-com...@gnu.org > Date: Sun, 16 Feb 2014 07:28:09 +0000 > Sender: groff-commit-bounces+wl=vcs.savannah.gnu....@gnu.org > Reply-To: Ingo Schwarze <schwa...@usta.de> Note: Sender: groff-commit-bounces+wl=vcs.savannah.gnu....@gnu.org I think that is the normal one generated on vcs. The mailing list only records one message sent for that commit. http://lists.gnu.org/archive/html/groff-commit/2014-02/threads.html I think that for any regular subscriber they would have only gotten one copy of the message. I think you are special. :-) I can only imagine that there is an alias somewhere for groff-commit that is coming to you directly and that is why you have received a second one. I looked but could not find an alias anywhere however. Not sure why you are getting a second copy. Could you forward two messages with all of the Received headers intact so that we could trace through the mail relays and see where it came through? > Additionally, the `Reply-To' field is set to the author of the patch. > However, I would prefer if it was set to `gr...@gnu.org', too. The Mail-Followup-To: gr...@gnu.org header is insufficient? But okay. Have adjusted the configuration so that the reply-to is set to gr...@gnu.org. Since groff-commit isn't a discussion list. Since I was looking and touching things I gratuitiously removed the duplicated subject tag. Personally I don't like subject tags at all, filing should be done with List-id:, and seeing 22 characters consumed from the subject that way was just too much. > Finally, while doing `git push', I see > > remote: *** no recipients configured so no email will be sent > remote: *** for 'refs/heads/master' update \ > c82e9a5ae45ca06ea95fa22b0f7fd37ad0e64be6->6d9c6728213e047bccd622bfb0 Hmm... That does seem egregious since the committer has no control over this configuration. Maybe setting it to "nobody" would be kinder since then it would see a recipient and would avoid that message. Except that for some reason on Savannah the nobody user forwards to root. And the alias list seems to be automatically generated by a process that I am not yet aware of. I will put that on the list to investigate. In the meantime I have hacked the git_multimail.py script to avoid that message. I agree that it isn't nice to send that to the committer. > None of those glitches are serious, but I want to report them > anyway :-) Maybe you want to forward them to the multimailhook people. Unfortunately I lack the time to take on an additional project at this point. I am still learning Savannah. I will just make a local modification to avoid the spurious commit warning message about no recipients. But if you or someone else wished to request new features of the upstream that would be great. https://github.com/mhagger/git-multimail Bob