From: Eric Sunshine <sunshine@sunshineco.com>
Date: Tue, 24 Nov 2020 17:24:57 GMT
Subject: Re: [PATCH 1/1] maintenance: fix a SEGFAULT when no repository
Message-ID: <CAPig+cQ-iWVz2Q1PtvbV0hk_HHRFqAFjxAF2DZ6doh2RxpZJhw@mail.gmail.com>
In-Reply-To: <20201124164405.29327-2-rafaeloliveira.cs@gmail.com>

On Tue, Nov 24, 2020 at 11:45 AM Rafael Silva
<rafaeloliveira.cs@gmail.com> wrote:
> The "git maintenance run" and "git maintenance start" commands holds a
> file-based lock at the .git/maintenance.lock and .git/schedule.lock
> respectively. These locks are used to ensure only one maintenance process
> is executed at the time as both operations involves writing data into
> the git repository.
>
> The path to the lock file is built using the "the_repository->objects->odb->path"
> that results in SEGFAULT when we have no repository available as
> "the_repository->objects->odb" is set to NULL.

This issue came up in review recently[1] in an unrelated way.

[1]: https://lore.kernel.org/git/CAPig+cRFQfg-NLx5dO+BjQpYduhOYs-_+ZRd=DhO8ebWjGB0iA@mail.gmail.com/

> Let's teach the maintenance_run_tasks() and update_background_schedule() to return
> an error and fails the command when we have no repository available.
>
> Signed-off-by: Rafael Silva <rafaeloliveira.cs@gmail.com>
> ---
> diff --git a/t/t7900-maintenance.sh b/t/t7900-maintenance.sh
> @@ -441,4 +441,9 @@ test_expect_success 'register preserves existing strategy' '
> +test_expect_success 'run and start command fails when no git repository' '
> +       test_must_fail git -C /tmp/ maintenance run &&
> +       test_must_fail git -C /tmp/ maintenance start
> +'

I wouldn't feel comfortable relying upon existence of /tmp/. It might
be sufficient to do this instead:

    mv .git save.git &&
    test_when_finished "mv save.git .git" &&
    test_must_fail git maintenance run &&
    test_must_fail git maintenance start