Re: [PATCH v15 05/13] ref-filter: implement an `align` atom

From: Matthieu Moy <matthieu.moy@grenoble-inp.fr>
Date: Sep 2, 2015, 15:50 UTC
Message-ID: <vpqsi6wom6q.fsf@anie.imag.fr>
In-Reply-To: <xmqqd1y0q30g.fsf@gitster.mtv.corp.google.com>

Junio C Hamano <gitster@pobox.com> writes:

Show 21 quoted lines

> Karthik Nayak <karthik.188@gmail.com> writes:
>
>>>> +             } else if (skip_prefix(name, "align", &valp)) {
>>>
>>> This looked as if you are willing to take %(align) in addition to
>>> %(align:...), but...
>>>
>>>> +                     struct align *align = &v->align;
>>>> +                     struct strbuf **s;
>>>> +
>>>> +                     if (valp[0] != ':')
>>>> +                             die(_("format: usage %%(align:<width>,<position>)"));
>>>
>>> ... apparently that is not what is happening.  Why not skip "align:"
>>> with colon as the prefix, then?
>>
>> Cause we wanted to provide an error for usage of "%(ailgn)" without any
>> subvalues as such.
>
> Wouldn't it be something that would be caught in the same codepath
> as what catches %(unrecognized) in the format string?

After thinking about it, I agree with Karthik: if we get the same codepath to complain about %(nosuchatom) and %(align), then we won't be able to provide an accurate error message. Or we would need to re-parse the atom, notice that it's one we know about, i.e. redo what we're already doing here.

-- 
Matthieu Moy
http://www-verimag.imag.fr/~moy/