Re: [PATCH] apply: don't use core.sharedRepository to create working tree files

From: Matheus Tavares Bernardino <matheus.bernardino@usp.br>
Date: Dec 22, 2020, 22:49 UTC
Message-ID: <CAHd-oW7XJL_a1zMAUetHzvrh8DrLT4g2awv-fjbTdeLVLKVsew@mail.gmail.com>
In-Reply-To: <CA+kUOamSd_3z8LbYt8QRx==HauYXoCe95B5hAW1W-LdnwGP-xw@mail.gmail.com>

On Tue, Dec 22, 2020 at 7:24 PM Adam Dinwoodie <adam@dinwoodie.org> wrote:

Show 22 quoted lines

>
> Cracked it, and it's a simple error in the test script. It wasn't
> readily obvious because the error gets silently swallowed, and
> presumably because the command isn't necessary on most *nix systems
> that have different behaviour for inheriting permissions, but the
> entire problem is fixed with the following diff:
>
> --- a/t/t4129-apply-samemode.sh
> +++ b/t/t4129-apply-samemode.sh
> @@ -78,7 +78,7 @@
>         test_config core.sharedRepository 0666 &&
>         (
>                 # Remove a default ACL if possible.
> -               (setfacl -k newdir 2>/dev/null || true) &&
> +               (setfacl -k . 2>/dev/null || true) &&
>                 umask 0077 &&
>
>                 # Test both files (f1) and leading dirs (d)
>
> It looks like the erroneous line was copied from t0001-init.sh, but
> that's a test where "newdir" is actually an existent directory, where
> we never use a directory of that name in this test script.

My bad, I should have been more careful here. Thanks for finding the problem!

> I'm assuming this is a simple and obvious enough fix that it can just
> get squashed into the original commit, but I don't know if that breaks
> things given the original commit is now included in rc tags. Let me
> know if I need to format and submit this as a full patch?

Yeah, since the original patch is already merged into `master`, I think a new patch fixing the problem would be more appropriate.

Thanks, Matheus