From: Junio C Hamano <gitster@pobox.com>
Date: Tue, 11 Aug 2020 20:28:22 GMT
Subject: Re: What's cooking in git.git (Aug 2020, #02; Mon, 10)
Message-ID: <xmqqy2ml7wrd.fsf@gitster.c.googlers.com>
In-Reply-To: <87364t563r.fsf@osv.gnss.ru>

Sergey Organov <sorganov@gmail.com> writes:

> And I asked you to postpone that in <875z9v7b0k.fsf@osv.gnss.ru> for
> which I got no answer. I repeat it here for convenience:
>
> "
> For now, can we leave these patches as-is and then add additional tests
> on top, please? I'm afraid that adding them now will force me to rewrite
> current descriptions, as the result won't be minimal anymore, and I
> feel uneasy about writing commit messages for these permuted tests
> anyway.
> "

I was expecting just one case I mentioned in the first message and
nothing else, which is crucial to highlight how --first-parent's
"implication" works with explicit --diff-merges.

Full "permutated tests" was what you brought up in the discussion,
but I already said it was overkill to your response before you said
the above.  The "minimal" would be the posted patch plus a missing
"first-parent comes before --diff-merges=off" test filled in.