From: Junio C Hamano <gitster@pobox.com>
Date: Thu, 23 Aug 2012 17:51:42 GMT
Subject: Re: [PATCH] specifying ranges: we did not mean to make ".." an empty set
Message-ID: <7vobm14hu9.fsf@alter.siamese.dyndns.org>
In-Reply-To: <874nntq0sb.fsf@thomas.inf.ethz.ch>

Thomas Rast <trast@student.ethz.ch> writes:

> I don't think that ".." is really a no-op.  It is true that HEAD..HEAD
> does not itself result in any revisions, but it *could* be used as a
> silly shorthand to introduce ^HEAD into the objects being walked.  This
> can make a difference if it then excludes other objects, too.
>
> I would argue that such use is misguided, and I am in favor of the
> patch, but in theory it is possible.

You could say "log --all .. --" when you mean "log --all ^HEAD",
because .. === HEAD..HEAD === ^HEAD HEAD and HEAD is already
contained in --all.  So it is a valid point.

You however cannot say "log --all .." without double-dash to
disambiguate, as you would get:

    $ git log --all ..
    fatal: ambiguous argument '..': both revision and filename

Any existing practice that used to produce useful results couldn't
have been using ".." without an explicit "--", I think, and with the
disambiguation that favors to take ".." as the parent directory,
"log --all .. --" still means "log --all ^HEAD".

So I think we are still OK.