Re: [PATCH 2/2] git-p4: Fix copy detection test

From: Junio C Hamano <gitster@pobox.com>
Date: Mar 28, 2015, 16:12 UTC
Message-ID: <xmqq619lyub8.fsf@gitster.dls.corp.google.com>
In-Reply-To: <20150328003636.07930f83@pt-vhugo>

Vitor Antunes <vitor.hda@gmail.com> writes:

Show 17 quoted lines

> Vitor Antunes <vitor.hda@gmail.com> wrote:
>
>>Junio C Hamano <gitster@pobox.com> wrote:
>>>Pete, these tests blame to your 9b6513ac (git p4 test: split up big
>>>t9800 test, 2012-06-27).  I presume that you tested the result of
>>>this splitting, but do you happen to know if we did something to
>>>cause the test to break recently?
>>
>>I also worked on these tests at that time and they were passing before and
>>after the reorganization. I'll prepare a bisect script and will try to find the
>>commit that started making this test fail.
>
> According to bisect, this is the first commit that makes the test fail:
>
> 7c85f8acb2282e3ed108c46b59fd5daa78bf17db
>
> Does this make sense to you?

Yeah, as the blamed commit changes the way the hashtable is used record and choose the rename source candidates, it is not surprising if it changes how two or more candidates with the same rename score are tie-broken.

Thanks.