Re: [PATCH v2] test_cmp: diagnose incorrect arguments

From: Eric Sunshine <sunshine@sunshineco.com>
Date: Aug 9, 2020, 19:34 UTC
Message-ID: <CAPig+cTA_BVz-bGcGY++g2RV-Zd3Gv_yTGjv=VeMeqhP4eZZrw@mail.gmail.com>
In-Reply-To: <xmqq7du7eiqi.fsf@gitster.c.googlers.com>

On Sun, Aug 9, 2020 at 3:12 PM Junio C Hamano <gitster@pobox.com> wrote:

Show 7 quoted lines

> The motivation makes quite a lot of sense.
>
> I suspect that it is a bug to use "-", i.e. "read from the standard
> input", for "$2", because
>
>  (1) if it is used for the expected output, we got the comparison in
>      the wrong direction;

I had the same concern that it didn't necessarily make sense to allow "-" for $2, but eventually thought of a case such as:

    sed '...' actual | test_cmp expect - &&

which massages 'actual' before test_cmp() sees it. True, we don't actually have such callers, but it theoretically is legitimate.

>  (2) if it is used for the actual output, we have a command whose
>      output we are interested in on the upstream side of a pipe to
>      discard its exit status.

If the command upstream is a Git command, that could indeed be a reason for concern, but, as illustrated above, it could just be a system command.