From: Junio C Hamano <gitster@pobox.com>
Date: Wed, 10 Sep 2008 00:02:09 GMT
Subject: Re: [PATCH] t7501: always use test_cmp instead of diff
Message-ID: <7vljy03m2m.fsf@gitster.siamese.dyndns.org>
In-Reply-To: <20080909235414.GA4768@coredump.intra.peff.net>

Jeff King <peff@peff.net> writes:

> On Wed, Sep 10, 2008 at 01:41:06AM +0200, Miklos Vajna wrote:
>
>>  test_expect_success \
>>      'validate git-rev-list output.' \
>> -    'diff current expected'
>> +    'test_cmp current expected'
>
> We seem to use the convention of
>
>   test_cmp <expected> <actual>
>
> elsewhere, rather than
>
>   test_cmp <actual> <expected>
>
> as you have here.  Most noticeably, that means the diff will show
> deviations from expected, rather "what should be done to make this as
> expected".

Yes, please.  I am glad to see somebody noticed it (I've been fixing these
whenever I touched vicinity of the ones that had comparisons swapped).