Re: no-xmailer tests fail under Mac OS

From: Michael Blume <blume.mike@gmail.com>
Date: Dec 6, 2014, 06:27 UTC
Message-ID: <CAO2U3QjkJRPrFvapK=LDU_GG5Ne6_zVD+S61JP+iV_Xi-gk5JQ@mail.gmail.com>
In-Reply-To: <20141206053434.GF31301@peff.net>

On Fri, Dec 5, 2014 at 9:34 PM, Jeff King <peff@peff.net> wrote:

Show 35 quoted lines

> On Fri, Dec 05, 2014 at 06:05:24PM -0800, Michael Blume wrote:
>
>> Failures start from
>>
>> commit d2384abff7a6181fd7b9a51af7e780aa21e5cb8d (refs/bisect/bad)
>> Author: Luis Henriques <henrix@camandro.org>
>> Date:   Thu Dec 4 19:11:30 2014 +0000
>>
>>     test/send-email: --[no-]xmailer tests
>>
>>     Add tests for the --[no-]xmailer option.
>>
>>     Signed-off-by: Luis Henriques <henrix@camandro.org>
>>     Signed-off-by: Junio C Hamano <gitster@pobox.com>
>>
>> but continue with Junio's SQUASH??? commit at b728d078
>
> The commit contains:
>
>   +       test "z$(grep ^X-Mailer: out | wc -l)" = "z$expected"
>
> We have had trouble in the past with "wc -l" output not being strictly
> portable. I do not recall offhand which systems, but it is a good bet
> that this is the culprit. Doing:
>
>   grep ^X-Mailer: out >mailer &&
>   test_line_count = $expected mailer
>
> should fix it. It might be even nicer to actually compare the x-mailer
> line we find to an expected output, but that may introduce complications
> if the value changes with the version or something (you'd have to
> sanitize the output, and then I do not know that the test is really
> buying much over just seeing whether it exists).
>
> -Peff

Actually need to drop the '&&', but yes, that works perfectly, thanks =)