Re: [PATCH 16/18] entry.c: update submodules when interesting

From: Junio C Hamano <gitster@pobox.com>
Date: Mar 8, 2017, 01:08 UTC
Message-ID: <xmqqpohs4ltc.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <CAGZ79kYmaAY-ASa672eXmQapxJns+a3dEC_BiEmH7ti==arLtw@mail.gmail.com>

Stefan Beller <sbeller@google.com> writes:

Show 6 quoted lines

>> In addition to mkdir(), would we also need the .git file that point
>> into superproject's .git/modules/ too?
>
> The move_head function takes care of it
> Both creation as well as deletion are handled in the move_head function,
> when either new or old is NULL.

Oh, if it does the creation of directory and adjusting of .git gitfile, and handles other anomalies, I have no problem with it.

It was just that this codepath tried to handle some kind of anomaly (namely, the user originally thought the submodule needs to be changed to a regular file and then changed mind and wants to restore it from the index) and yet not doing the full anomaly handling (like "not even an empty directory exists" case) that confused me.