From: Junio C Hamano <gitster@pobox.com>
Date: Fri, 01 Jul 2016 18:54:04 GMT
Subject: Re: [PATCH v2 00/12] nd/icase updates
Message-ID: <xmqqziq1xk8j.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <CACsJy8BBQqxzT8jdaysyeDv3ja=DTD6R-whA3Lhr3pEs-2x8dA@mail.gmail.com>

Duy Nguyen <pclouds@gmail.com> writes:

> On Fri, Jul 1, 2016 at 8:18 PM, Junio C Hamano <gitster@pobox.com> wrote:
>> Duy Nguyen <pclouds@gmail.com> writes:
>>
>>> Makes sense. But then if opt->ignore_case is false and regflags
>>> happens to have REG_ICASE set, should we clear it as well?
>>
>> I think .ignore_case is set iff '-i' is given, and .regflags has
>> REG_ICASE only if '-i' is given and the user said she does not want
>> literal string match (i.e. no '-F').
>>
>> So... can .regflags have REG_ICASE when .ignore_case is false?
>
> Yeah reg_icase is more like a subset of ignore_case. Ignore what I wrote.

The rule of the game can be updated to make whoever looks at
.ignore_case currently instead look at .regflags & REG_ICASE; that
way, we could remove .ignore_case and stop worrying about one is
subset of the other.

It is just it felt funny to be mucking with .regflags when the user
explicitly said -F, and that is why they are separate.