Re: [PATCH] t4015: let the test pass with any default branch name

[Johannes Schindelin <Johannes.Schindelin@gmx.de>]

Hi Junio,

On Wed, 18 Nov 2020, Junio C Hamano wrote:

> "Johannes Schindelin via GitGitGadget" <gitgitgadget@gmail.com>
> writes:
>
> > From: Johannes Schindelin <johannes.schindelin@gmx.de>
> >
> > We do not need to hard-code the actual branch name, as we have access to
> > the reflog and can be a lot more precise at what we want.
>
> I am not enthused to see us going beyond "@{1}" or "@{-1}".

Right. It also makes it a bit harder to understand the test case for the
occasional reader.

> It would make the test a bit too brittle as a future changes need to
> refrain from creating another commit and/or switching to a third branch
> to do some more set-up before coming here to attempt an merge.
>
> In this particular case, "git merge @{-1}" may be more robust from
> that point of view

It might be _slightly_ more robust, but not by a lot. I opted to use
`test_commit` instead and then refer to the revision via the tag that is
created by that function.

Ciao,
Dscho

> > diff --git a/t/t4015-diff-whitespace.sh b/t/t4015-diff-whitespace.sh
> > index 8bdaa0a693..c9dd62b70e 100755
> > --- a/t/t4015-diff-whitespace.sh
> > +++ b/t/t4015-diff-whitespace.sh
> > @@ -883,7 +883,8 @@ test_expect_success 'combined diff with autocrlf conversion' '
> >  	echo >x goodbye &&
> >  	git commit -m "the other side" x &&
> >  	git config core.autocrlf true &&
> > -	test_must_fail git merge master &&
> > +	test_must_fail git merge @{2} >actual &&
> > +	test_i18ngrep "Automatic merge failed" actual &&
> >
> >  	git diff >actual.raw &&
> >  	sed -e "1,/^@@@/d" actual.raw >actual &&
> >
> > base-commit: e31aba42fb12bdeb0f850829e008e1e3f43af500
>
>