From: Santi Béjar <santi@agolina.net>
Date: Fri, 03 Dec 2010 22:16:11 GMT
Subject: Re: BUG: git pull --rebase on detatched head prints a lot of usage warnings
Message-ID: <AANLkTimgk-jqyKD+ab=bHZLVH3F52WqCjr4v3chMY_Fv@mail.gmail.com>
In-Reply-To: <7vbp52o9c6.fsf@alter.siamese.dyndns.org>

On Fri, Dec 3, 2010 at 10:06 PM, Junio C Hamano <gitster@pobox.com> wrote:
> Santi Béjar <santi@agolina.net> writes:
>
>> I also think it fixes, but I wonder if it's the best fix. One thing is
>> that show-branch is an ancillary command, and not a plumbing command
>> (but there is already a call to show-branch, but it should probably be
>> change to merge-base now that it handles three commit as arguments).
>
> Are you sure "show-branch --merge-base A B C" and "merge-base A B C" are
> meant to compute the same thing?

No, I'm not sure. In fact the documentation explicit says that they
are different. But I fail to see when it matters, and more
specifically (now that I've reread the docs), what is the point of
"merge-base A B C".

Just a data point, show-branch --merge-base is only used once in the
code, in git-pull.sh, but "git merge-base A B C" is not used (at least
using my grep).

Or the equivalent is the undocumented "git merge-base --octopus A B C"?

>
> I do not think ancillary necessarily means Porcelain, but that is a
> separate issue.

Me neither. But is the promise to not change the behavior/output as
hard as with plumbing?

Santi