Re: [PATCH v2 2/2] attr: more matching optimizations from .gitignore

[Junio C Hamano <gitster@pobox.com>]

Nguyễn Thái Ngọc Duy  <pclouds@gmail.com> writes:

> diff --git a/t/t0003-attributes.sh b/t/t0003-attributes.sh
> index 51f3045..4a1402f 100755
> --- a/t/t0003-attributes.sh
> +++ b/t/t0003-attributes.sh
> @@ -242,4 +242,18 @@ test_expect_success 'bare repository: test info/attributes' '
>  	attr_check subdir/a/i unspecified
>  '
>  
> +test_expect_success 'leave bare' '
> +	cd ..
> +'
> +
> +test_expect_success 'negative patterns' '
> +	echo "!f test=bar" >.gitattributes &&
> +	test_must_fail git check-attr test -- f
> +'
> +
> +test_expect_success 'patterns starting with exclamation' '
> +	echo "\!f test=foo" >.gitattributes &&
> +	attr_check "!f" foo
> +'
> +
>  test_done

This is not entirely your fault, but please don't do that "cd ..".

The original test had "cd bare", made an assumption that step will
never fail (which is mostly correct), and ran everything afterward
in that subdirectory.

Adding "Do a 'cd ..' to come back" is a horrible way to build on
it.  Imagine what happens when another person also did the same
thing, and both changes need to be merged.  You will end up going up
two levels, which is not what you want.

I think the right fix is to make each of the test that wants to run
in "bare" chdir in its own subshell, or not append these new tests
that do not run in the "bare" to the end of this file, but before
the execution goes down to "bare".