Re: [PATCH 1/2] t/t1417: test symbolic-ref effects on ref logs

From: Junio C Hamano <gitster@pobox.com>
Date: Jan 31, 2021, 00:11 UTC
Message-ID: <xmqqo8h6nexk.fsf@gitster.c.googlers.com>
In-Reply-To: <81AEEED6-26EC-4F32-AA65-B028435D812D@gmail.com>

"Kyle J. McKay" <mackyle@gmail.com> writes:

> I'm having a bit of trouble parsing that into expectations.  A little
> help please.
> Are you suggesting that (1) just be omitted?  Or that it be modified
> so that it's an "expect success" patch?

Neither.

The result of applying the current 1/2 and 2/2 on top of, say 'master', would be the shape of the tree you would want to be in.

Our preference is just to have it as a single patch, not as "first expect failure and then flip it to expect success while modifying the code". That approach makes the second step harder to review than necessary, because the "git show" output and "format-patch" output from the step would show only very little about the test that changes behaviour.

Even with a single patch, if somebody wants a demonstration of what used to be broken without the code modification, it is easy to apply only the test part of the single patch without using the code change to see how it breaks, so "I want to demonstrate the breakage" is not a reason to have it as a separate step.

Thanks.