Re: [PATCH v4] apply.c: fix -p argument parsing

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

Mirko Faina <mroik@delayed.space> writes:

> As Jeff pointed out, the previous patch doesn't pass tests on windows...
> Inlined as a workaround and to avoid adding additional folders to the
> existing test directory.

Thanks for working very well together.
> +test_expect_success 'git apply -p 1 patch' '
> +	cat >patch <<-\EOF &&
> +		From 90ad11d5b2d437e82d4d992f72fb44c2227798b5 Mon Sep 17 00:00:00 2001
> +		From: Mroik <mroik@delayed.space>
> +		Date: Mon, 9 Mar 2026 23:25:00 +0100
> +		Subject: [PATCH] Test
> +
> +		---
> +		 t/test/test | 0
> +		 1 file changed, 0 insertions(+), 0 deletions(-)
> +		 create mode 100644 t/test/test
> +
> +		diff --git a/t/test/test b/t/test/test
> +		new file mode 100644
> +		index 0000000000..e69de29bb2
> +		-- 
> +		2.53.0.851.ga537e3e6e9
> +
> +	EOF

It is more customary to indent the here-doc body to the same level
as surrounding <<EOF..EOF; no need to resend only to fix this, as I
can easily dedent it by one level.

> +	test_when_finished "rm -rf t" &&
> +	git apply -p 1 patch &&
> +	test_path_is_dir t
> +'
> +
> +test_expect_success 'apply fails due to non-num -p' '
> +	test_when_finished "rm -rf t test" &&
> +	test_must_fail git apply -p malformed patch
> +'
> +
> +test_expect_success 'apply fails due to trailing non-digit in -p' '
> +	test_when_finished "rm -rf t test" &&
> +	test_must_fail git apply -p 2q patch
> +'
> +
> +test_expect_success 'apply fails due to negative number in -p' '
> +	test_when_finished "rm -rf t test patch" &&
> +	test_must_fail git apply -p -1 patch
> +'

The test all make sense, but if we know what error message we are
expecting, it may not be a bad idea to do something like

	test_must_fail git apply -p -1 patch 2>err &&
	test_grep "<num> has to be a non-negative" err

to ensure that the command did not fail for a wrong reason.

THanks.