Re: [PATCH v2 02/14] pull: improve default warning

From: Junio C Hamano <gitster@pobox.com>
To: Felipe Contreras <felipe.contreras@gmail.com>
Cc: "Jacob Keller" <jacob.keller@gmail.com>,
	"Elijah Newren" <newren@gmail.com>,
	"Git Mailing List" <git@vger.kernel.org>,
	"Vít Ondruch" <vondruch@redhat.com>,
	"Alex Henrie" <alexhenrie24@gmail.com>,
	"Jeff King" <peff@peff.net>,
	"Philip Oakley" <philipoakley@iee.email>,
	"Theodore Y . Ts'o" <tytso@mit.edu>,
	"John Keeping" <john@keeping.me.uk>,
	"Richard Hansen" <rhansen@rhansen.org>,
	"Brian M. Carlson" <sandals@crustytoothpaste.net>,
	"W. Trevor King" <wking@tremily.us>
Subject: Re: [PATCH v2 02/14] pull: improve default warning
Date: Tue, 08 Dec 2020 12:16:46 -0800	[thread overview]
Message-ID: <xmqqy2i86ok1.fsf@gitster.c.googlers.com> (raw)
In-Reply-To: <CAMP44s3KCoDfRXzarJw5AE7UsY-=eP6GbHzdDcdrs2rsw5tL+w@mail.gmail.com> (Felipe Contreras's message of "Mon, 7 Dec 2020 21:15:05 -0600")

Felipe Contreras <felipe.contreras@gmail.com> writes:

> That is exemplified by the fact that this whole thread started from a
> user that refused to configure pull.rebase and expected the Git
> project to just do the right thing (which is basically choosing a
> useful default).

Which is basically asking for impossible, and I do not think it is a
good idea to focus our effort to satisfy such a request in general.
There is no useful default that suits everybody in this particular
case.

The "force ff-only" approach indeed gives a very sensible default
behaviour of dying for those who haven't expressed the choice
between rebase and merge in a situation where the difference between
the two results in histories of different shapes.

But for anybody who uses git for real (read: produces his or her own
history), it would be pretty much a useless default that forces the
user to say rebase or merge every time 'git pull' is run.  That is
why I am not enthused by the pull.mode=(rebase/merge/ff-only)
configuration.  The third choice does help completeness.  When a
user asks "the documentation tells pull.mode can be set to
non-default behaviour---what value can one set it to to get the
default behaviour of not allowing any original work?", it can be
answered if we had pull.mode=ff-only.  But other than that, I do not
see any real use for the choice, which would mean in practice,
pull.mode would have only two useful values, rebase or merge.  That
does not feel a good enough reason to supersede what already exists,
which is pull.rebase=yes/no.

Perhaps there is a good reason why certain classes of users would
want to configure pull.mode=ff-only (i.e. "I might type 'git pull'
by mistake, please stop me if I did so on a branch I have real work
already.").  If that is the case, I would very much agree that it
would be awkward to express that choice in the current framework to
choose between pull.rebase=yes/no and pull.mode=(rebase/merge/ff-only)
would become a lot easier to explain.

I dunno.