Re: [PATCH v3 2/2] submodule: don't print status output with ignore=all

From: Junio C Hamano <gitster@pobox.com>
Date: Sep 4, 2013, 20:57 UTC
Message-ID: <xmqqli3cl2yk.fsf@gitster.dls.corp.google.com>
In-Reply-To: <522791AE.8070103@web.de>

Jens Lehmann <Jens.Lehmann@web.de> writes:

Show 42 quoted lines

> Am 03.09.2013 21:53, schrieb Junio C Hamano:
>> Jens, I see 1/2 is the same as the previous one you already acked.
>
> Yep.
>
>> Does this update to 2/2 look good to you?  Sorry, but I lost track
>> of the discussion that led to this reroll, hence a ping.
>
> v3 fixes the bug Matthieu noticed, I only had some remarks to the
> new test Brian added. If you could replace his patch to t7401 with
> the following diff it's an ack from me on this one too.
>
> -------------------------8<---------------------------
> diff --git a/t/t7401-submodule-summary.sh b/t/t7401-submodule-summary.sh
> index ac2434c..81ae7c9 100755
> --- a/t/t7401-submodule-summary.sh
> +++ b/t/t7401-submodule-summary.sh
> @@ -104,6 +104,24 @@ EOF
>  	test_cmp expected actual
>  "
>
> +test_expect_success 'no ignore=all setting has any effect' "
> +	git config -f .gitmodules submodule.sm1.path sm1 &&
> +	git config -f .gitmodules submodule.sm1.ignore all &&
> +	git config submodule.sm1.ignore all &&
> +	git config diff.ignoreSubmodules all &&
> +	git submodule summary >actual &&
> +	cat >expected <<-EOF &&
> +* sm1 $head1...$head2 (1):
> +  > Add foo3
> +
> +EOF
> +	test_cmp expected actual &&
> +	git config --unset diff.ignoreSubmodules &&
> +	git config --remove-section submodule.sm1 &&
> +	git config -f .gitmodules --remove-section submodule.sm1
> +"
> +
> +
>  commit_file sm1 &&
>  head3=$(
>  	cd sm1 &&

Thanks.

The above patch makes the <<-EOF situation that already exists in this script worse. The only reason we would say -EOF not EOF is because we would want to indent the here-document to align with the rest of the command sequence, so we should either indent with HT, or drop the dash. I suspect the original did it that way fearing that someday the indentation of the submodule difference list might start using HT, but I do not think that is likely to happen, so my vote goes to keeping '-' and indenting.

We need a clean-up patch after this series settles.