Re: [PATCH v2 1/4] builtin: add git-check-mailmap command

From: Eric Sunshine <sunshine@sunshineco.com>
Date: Jul 12, 2013, 06:39 UTC
Message-ID: <CAPig+cQz8jAHw3p4Fi30-KGZoVcb1vkgS9yA23abJVJLpDrvUA@mail.gmail.com>
In-Reply-To: <7vmwpsnv5l.fsf@alter.siamese.dyndns.org>

On Fri, Jul 12, 2013 at 2:34 AM, Junio C Hamano <gitster@pobox.com> wrote:

Show 9 quoted lines

> Eric Sunshine <sunshine@sunshineco.com> writes:
>
>> ... Is it desirable to do so
>> or should the user have more fine-grained control? ("xargs -0" comes
>> to mind when thinking of a null-termination input switch.)
>
> For the purposes of check-attr and check-ignore, a single "-z"
> governing both is sufficient.  I think you already got that from my
> 4-patch series, but the core reason for that is :

I'm reading it right now.

Show 12 quoted lines

>  - when "-z" is used, the user knows the input paths may need
>    protection against LF.
>
>  - our output contains these same paths.
>
>  - which means our output cannot be expressed unambiguously using LF
>    as record separator.
>
> For the purpose of check-mailmap, I actually do not see much point
> in supporting "-z" format.  We do not even handle names or addresses
> with LF in it.  The mailmap format would not let you express such
> records in the first place, no?

You're right. I had this exact argument in mind for why null-termination was not needed on the input side of check-mailmap, but for some reason I had a blind spot concerning the output side. I'll drop this option in the next re-roll.