From: D. Ben Knoble <ben.knoble@gmail.com>
Date: Wed, 18 Mar 2026 16:36:04 GMT
Subject: Re: [PATCH] apply: fix new-style empty context line triggering incomplete-line check
Message-ID: <CALnO6CDNwa8Ez4Ug0f8zNyxF1n3C_j8mLRbH7wChVioNoC5QVw@mail.gmail.com>
In-Reply-To: <xmqqcy12l2ft.fsf@gitster.g>

On Tue, Mar 17, 2026 at 2:48 PM Junio C Hamano <gitster@pobox.com> wrote:
>
> Eric Sunshine <sunshine@sunshineco.com> writes:
>
> >> +       test_when_finished "rm -f sample*-i patch patch-new target" &&
> >> +       (test_write_lines 1 2 3 "" 4 5 ) >sample-i &&
> >> +       (test_write_lines 1 2 3 "" 0 5 ) >sample2-i &&
> >
> > Curious. Why are the `test_write_line` invocations wrapped in parentheses?
> >
> > Also, is the whitespace before the closing parenthesis intentional?
> >
> >>  test_expect_success 'incomplete context line (not an error)' '
> >>         (test_write_lines 1 2 3 4 5 && printf 6) >sample-i &&
> >>         (test_write_lines 1 2 3 0 5 && printf 6) >sample2-i &&
> >
> > Perhaps the parentheses in the new test were copied from some existing
> > test, such as this, which already used them for a legitimate reason?
>
> Yes, the existing one was concatenating output from two commands run
> in a row into a single redirection, so (grouping of the commands) in
> parentheses were justifiable.
>
> The new one does not have such a justification.  Thanks for
> noticing.

I think braces { test_writes lines … && printf … ; } would have
sufficed for the second example, and might be cheaper (avoiding the
extra process for the subshell, which we've been told is especially
expensive for Windows).

-- 
D. Ben Knoble