From: Junio C Hamano <gitster@pobox.com>
Date: Wed, 03 Jun 2015 18:25:38 GMT
Subject: Re: [RFC/PATCH 2/2] status: fix tests to handle new verbose git status during rebase
Message-ID: <xmqqvbf4wswt.fsf@gitster.dls.corp.google.com>
In-Reply-To: <vpqmw0gadth.fsf@anie.imag.fr>

Matthieu Moy <Matthieu.Moy@grenoble-inp.fr> writes:

> Guillaume Pagès <guillaume.pages@ensimag.grenoble-inp.fr> writes:
>
>> Signed-off-by: Guillaume Pagès <guillaume.pages@ensimag.grenoble-inp.fr>
>> ---
>>  t/t7512-status-help.sh | 227 ++++++++++++++++++++++++++++++++++++++++++++++---
>
> Your history is not bisectable if you sparate tests and code changes in
> two patches.

Yes.

And by squashing the two, you do not have to label 2/2 incorrectly
as "fix tests"; it is merely fixing what 1/2 broke ;-)

> Plus, as a reviewer, I like seeing changes to the tests next to changes
> to the code, to show me what is the impact of the code change on the
> output of the program.

That too. is a very important consideration, maybe even more so than
bisectability.

Thanks.