Re: [PATCH v2 3/4] t4203: test mailmap functionality directly rather than indirectly

From: Junio C Hamano <gitster@pobox.com>
Date: Jul 12, 2013, 06:25 UTC
Message-ID: <7vr4f4nvkw.fsf@alter.siamese.dyndns.org>
In-Reply-To: <CAPig+cSfqONOFmmXX=s+HWKnLr5eG27jU9xgE7Dki0sycPPz1g@mail.gmail.com>

Eric Sunshine <sunshine@sunshineco.com> writes:

Show 13 quoted lines

> On Fri, Jul 12, 2013 at 1:48 AM, Junio C Hamano <gitster@pobox.com> wrote:
>> Jonathan Nieder <jrnieder@gmail.com> writes:
>>
>>> My current thinking is "no" --- the patch has as a justification "Now
>>> we can test these aspects of .mailmap handling directly with a
>>> low-level tool instead of using the tool most people will use, so do
>>> so", which sounds an awful lot like "Reduce test coverage of commonly
>>> used tools, because we can".
>>
>> Yes, that was exactly my reaction that prompted my response.
>
> Does any of my follow-up commentary result in a different
> reaction?

Not really. While I _do_ think direct testing has merits, I think that should be done by adding direct tests, not by removing the tests that are meant to protect higher level _users_ of the underlying mechanism from breakages. After all, their breakages may not come from new breakages of the lower level mechanism (i.e. the mailmap.c code) but the way these higher level code makes calls into the mechanism, and direct test of the lower level mechanism will not protect them from the latter kind of breakages.