From: Karthik Nayak <karthik.188@gmail.com>
Date: Mon, 27 Jul 2015 18:54:33 GMT
Subject: Re: [PATCH v5 03/11] ref-filter: add option to pad atoms to the right
Message-ID: <CAOLa=ZT9_0ZZQwKwObgndLezcOWOk3LBTjAM7mMU2cyXu7--MQ@mail.gmail.com>
In-Reply-To: <vpqwpxl30aj.fsf@anie.imag.fr>

On Tue, Jul 28, 2015 at 12:17 AM, Matthieu Moy
<Matthieu.Moy@grenoble-inp.fr> wrote:
> Karthik Nayak <karthik.188@gmail.com> writes:
>
>> On Mon, Jul 27, 2015 at 9:24 PM, Matthieu Moy
>> <Matthieu.Moy@grenoble-inp.fr> wrote:
>>> Yes, but on the other hand we already have:
>>>
>>>   git log --format='%<|(50)A very long irrevlevancy|%an|'
>>>
>>> that pads/truncate %an. So, consistancy would dictate that Karthik's
>>> version is the right one.
>>
>> Sorry but I didn't understand what you're trying to say here, Matthieu.
>
> The "git log" equivalent of %(padright:N) is %<|(N), and it behaves the
> same way as your current implementation of %(padright) (except for the
> missing reset in your v5).
>
> So, if we want to be consistant with "git log", we should keep the
> "apply to next atom, even if it's far away in the format string"
> semantics.
>
> Note that consistancy is not the only criterion for choice, so I'm not
> saying we should absolutely do this, just that there's an argument in
> favor of it.
>

I didn't know that, thanks, I think I'll let Junio make the call here :)

-- 
Regards,
Karthik Nayak