Re: [PATCH] Propagate --quiet on submodule update to merge/rebase

From: Junio C Hamano <gitster@pobox.com>
Date: Sep 30, 2020, 19:24 UTC
Message-ID: <xmqqft6zgjaj.fsf@gitster.c.googlers.com>
In-Reply-To: <20200930074729.99629-1-tbodt@google.com>

Theodore Dubois <tbodt@google.com> writes:

Show 30 quoted lines

> Without this, commands such as
> git pull --rebase --recurse-submodules --quiet
> might produce non-quiet output from the merge or rebase.
>
> Signed-off-by: Theodore Dubois <tbodt@google.com>
> ---
>  git-submodule.sh            | 4 ++--
>  t/t7406-submodule-update.sh | 9 +++++++++
>  2 files changed, 11 insertions(+), 2 deletions(-)
>
> diff --git git-submodule.sh git-submodule.sh
> index 6fb12585cb..5c22b17221 100755
> --- git-submodule.sh
> +++ git-submodule.sh
> @@ -614,13 +614,13 @@ cmd_update()
>  				say_msg="$(eval_gettext "Submodule path '\$displaypath': checked out '\$sha1'")"
>  				;;
>  			rebase)
> -				command="git rebase"
> +				command="git rebase ${GIT_QUIET:+--quiet}"
>  				die_msg="$(eval_gettext "Unable to rebase '\$sha1' in submodule path '\$displaypath'")"
>  				say_msg="$(eval_gettext "Submodule path '\$displaypath': rebased into '\$sha1'")"
>  				must_die_on_failure=yes
>  				;;
>  			merge)
> -				command="git merge"
> +				command="git merge ${GIT_QUIET:+--quiet}"
>  				die_msg="$(eval_gettext "Unable to merge '\$sha1' in submodule path '\$displaypath'")"
>  				say_msg="$(eval_gettext "Submodule path '\$displaypath': merged in '\$sha1'")"
>  				must_die_on_failure=yes

This is not the problem this patch introduces, but the way GIT_QUIET variable is set up does not allow us to do the above so nicely, I suspect. Wouldn't the above change make "git submodule update -v" invoke the underlying commands with "--quiet" option?

The problematic piece of code is this part:

        cmd_update()
        {
                # parse $args after "submodule ... update".
                while test $# -ne 0
                do
                        case "$1" in
                        -q|--quiet)
                                GIT_QUIET=1
                                ;;
                        -v)
                                GIT_QUIET=0
                                ;;
                        --progress)
                                progress=1
                                ;;

I think this is the only place in the script that GIT_QUIET is set to 0, but all the places that refer to the variable do not even check the value held in it. Makes me wonder if it was used differently back when e84c3cf3dc3 was written.

    ... goes and looks at the offending commit ...

I think the right fix could have been "unset GIT_QUIET" instead of assigning 0 that means the same thing as GIT_QUIET=1

In any case, the posted patch is a good first step but it makes the existing problem worse. Let's fix GIT_QUIET=0 at the same time.

Thanks.

Show 12 quoted lines

> diff --git t/t7406-submodule-update.sh t/t7406-submodule-update.sh
> index aa19ff3a2e..5213e47af8 100755
> --- t/t7406-submodule-update.sh
> +++ t/t7406-submodule-update.sh
> @@ -1022,4 +1022,13 @@ test_expect_success 'git clone passes the parallel jobs config on to submodules'
>  	rm -rf super4
>  '
>  
> +test_expect_success 'submodule update --quiet passes quietness to merge/rebase' '
> +	(cd super &&
> +	 test_commit -C rebasing message &&
> +	 git submodule update --rebase --quiet >out 2>err &&

IOW, I suspect that this test will still pass with s/--quiet/-v/ .

> +	 test_must_be_empty out &&
> +	 test_must_be_empty err
> +	)
> +'