From: Junio C Hamano <gitster@pobox.com>
Date: Mon, 16 Aug 2010 01:49:43 GMT
Subject: Re: [PATCH] log: test for regression introduced in v1.7.2-rc0~103^2~2
Message-ID: <7vaaonfhs8.fsf@alter.siamese.dyndns.org>
In-Reply-To: <AANLkTi=PAW_Owy_-DSQ32sboB28373Gb_aySbpeprwLg@mail.gmail.com>

Ævar Arnfjörð Bjarmason <avarab@gmail.com> writes:

>>> +     # Needs an unrelated root commit
>>> +     test_commit README &&
>>
>> This is not a "root" commit, is it?
>
> s/root/first/

It is not even the first commit, is it?  It comes on top of whatever
commits that earlier tests left.

>>> +     >Foo.bar &&
>>> +     git add Foo.bar &&
>>> +     git commit --allow-empty-message </dev/null &&
>>
>> Does emptiness of the message matter?
>
> No, I was just going for a minimal test case, no commit message is
> more minimal than having one.

I do not think having to write "--allow-empty-message </dev/null" is
aiming for being minimal; it is doing something unusual after all.

If you do not remember why you added this test 6 months down the road,
wouldn't you be confused to think maybe the commit has to be unusual in
that it has to lack the message to trigger the bug?