From: Junio C Hamano <gitster@pobox.com>
Date: Mon, 20 Jun 2016 15:53:16 GMT
Subject: Re: [PATCH v2 0/8] object_id part 4
Message-ID: <xmqq8txzvour.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <20160620100522.GB14058@sigill.intra.peff.net>

Jeff King <peff@peff.net> writes:

> I am on the fence regarding oidcpy/oidclr. I agree they _could_ be
> struct assignments, but it is also convenient to have concept wrapped up
> in a function, in case we ever want to do anything more complicated.

Also dedicated functions have documenation value.  There are some
things that currently use the same 40-hex that is the result of
running SHA-1 but are not object names (e.g.  patch id, and rerere
id).  They use the same hashcpy()/hashcmp() helpers as object names
do, but the code that use them probably do not want to be converted
to struct oid and oidcpy().