From: Matthieu Moy Date: Mon, 23 May 2016 20:05:55 GMT Subject: Re: [RFC-PATCH 2/2] t9001: adding --quote-mail option test Message-ID: In-Reply-To: <1464031829-6107-3-git-send-email-tom.russello@grenoble-inp.org> > Subject: [RFC-PATCH 2/2] t9001: adding --quote-mail option test We write messages at imperative tone, hence s/adding/add/ Tom Russello writes: > From: Tom Russello Please use the same identity for email and commit to avoid this line. > --- > > > diff --git a/t/t9001-send-email.sh b/t/t9001-send-email.sh No diffstat again? Splitting a patch series as "code first; tests after" is not a good idea IMHO. When questionning the behavior of To: Vs Cc: in the previous patch, I would have appreciated having tests in the same message, to check that the tested behavior was indeed the one I was reading in the code. OTOH, having one patch to introduce "--quote-email populates To: and Cc: headers", and then another one for "--quote-email quotes the message body" would make the review much easier. Oh, BTW, this obviously lacks documentation (Documentation/git-send-email.txt). And that's one reason why the diffstat is useful: one can reply "this lacks tests and doc" before even reviewing the patch. Regards, -- Matthieu Moy http://www-verimag.imag.fr/~moy/