Re: [PATCH v2 1/4] rev-parse: fix some options when executed from subpath of main tree

From: Junio C Hamano <gitster@pobox.com>
Date: May 6, 2016, 17:08 UTC
Message-ID: <xmqq8tzn5d26.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <20160506161312.Horde.7i8_sE5ISIqccneOIfinvCX@webmail.informatik.kit.edu>

SZEDER Gábor <szeder@ira.uka.de> writes:

>> I agree with this, but at one point Junio suggested that it should
>> return the relative path[1],
>
> I wasn't aware of Junio's suggestion.

Mike justified his position to change everything to absolute, citing that it is what "rev-parse --show-toplevel" does when run from a subdirectory. I just gave a counter-example that rev-parse does not always talk in absolute with "rev-parse -C t --show-cdup".

It hinted that his justification may not be strong enough to choose absolute over relative, but wasn't meant to be a suggestion to choose relative at all. At most, I was saying "either is equally plausible".

> It shouldn't really matter in practice, because both the absolute and
> relative paths will ultimately lead to the same place.  However, I
> still think that for consistency's sake absolute paths would be better
> when executed in a subdir of the working tree.

Sure.

One thing I'd be worried about is that people may be using it to compute paths and store them in $GIT_DIR/somewhere, and expecting that they can later do a wholesale "mv" of directory hierarchy and relative paths still work. I wouldn't be one of those who does the "mv" so I personally do not care, but I suspect some do.