Re: [RFC v2] git-multimail: a replacement for post-receive-email

Michael Haggerty [off-list ref] writes:

I wonder whether it would be to far off the beaten path to allow glob
patterns in the branch specification; e.g.,

   [multimailhook "refs/heads/release-*"]
         mailingList = qa@example.com

Yes, that would be even better.

For the case of multiple glob patterns matching a branch name, there
would probably have to be a notion of "best match", but that doesn't
seem too difficult.

I'd rather have a simple rule here like "last one wins" or so. Saying
that foo-bar-* is a better match than foo-* may be easy, but you can
hardly avoid having corner-cases like foo-*-boz vs foo-bar-* when
matching foo-bar-boz.

This feature could also be used to get the functionality of your
proposal for skipRefs and onlyRefs [1] in a more general way:

   [multimailhook]
         mailingList = some@example.com
   [multimailhook "refs/heads/user/$USER/*"]
         mailingList = ""

Yes, I thougth about that, but it is not only "more general", but also
"less conveinient":

    [multimailhook]
          mailingList = some@example.com
          refchangelist = other@example.com
    [multimailhook "refs/heads/user/$USER/*"]
          mailingList = ""
          # Oops, forgot to override refchangelist, the mail will still
          # be sent.

So skipRefs and onlyRefs would still make sense IMHO.

-- 
Matthieu Moy
http://www-verimag.imag.fr/~moy/