From: Junio C Hamano <gitster@pobox.com>
Date: Sat, 09 Nov 2019 04:27:16 GMT
Subject: Re: Change behavior of git add --patch on newly added file?
Message-ID: <xmqq5zjtn07v.fsf@gitster-ct.c.googlers.com>
In-Reply-To: <20191108225035.GA60198@google.com>

Emily Shaffer <emilyshaffer@google.com> writes:

> Should 'git add -p <newly-added-file>' do the same thing as 'git add -N
> <newly-added-file && git add -p <newly-added-file>'?

Probably.  

I originally wrote "git add -i" with the intention that the
interactive mode is _the_ primary interface to the machinery, so the
expected way to work with a new file was "git add -i", tell the
command to add that <newly-added-file>, and do the "patch" thing
using the interactive subcommand to do so within the "git add -i"
session.

Later people liked (only) the patch part, and "git add -p" (and
various "--patch" options that invoke "add -p" internally from other
commands like "checkout", "reset" were added) was born.  I think
nobody thought things through when they did so.

If I were designing "git add -p" from scratch and explicitly asked
not to do the other parts of the "--interactive" feature, I would
imagine "add -N && add -p" combination is what I would make it
mimic.

Patches welcome, but you may want to check with Dscho as there is an
effort going on to reimplement the entire "add -i" machinery in C.

Thanks.