Re: [PATCH v5 0/8] port the filtering part of branch.c to use ref-filter APIs

> On Mon, Sep 21, 2015 at 12:30 AM, Matthieu Moy
> <Matthieu.Moy@grenoble-inp.fr> wrote:
>> Karthik Nayak <karthik.188@gmail.com> writes:
>>
>>> --- a/t/t3203-branch-output.sh
>>> +++ b/t/t3203-branch-output.sh
>>> @@ -145,8 +145,8 @@ EOF
>>>
>>>  test_expect_success 'git branch `--sort` option' '
>>>       cat >expect <<-\EOF &&
>>> -     * (HEAD detached from fromtag)
>>>         branch-two
>>> +     * (HEAD detached from fromtag)
>>>         branch-one
>>>         master
>>>       EOF
>>
>> I thought this was obviously wrong, but this tests uses
>> --sort=objectsize, so it's OK. Indeed, HEAD and branch-one have
>> objectsize=215, and branch-two has objectsize=167.
>>
>> So, the previous code was hardcoding "detached HEAD always comes first",
>> but we now sort the whole list, including detached HEAD, so branch-two
>> comes first.
>>
>
> Yeah! Even Junio said this was better, and even I think this is what
> the user would expect.