From: Junio C Hamano <gitster@pobox.com>
Date: Tue, 10 Feb 2009 05:16:23 GMT
Subject: Re: [PATCH] Generalize and libify index_is_dirty() to index_differs_from(...)
Message-ID: <7v63jirhw8.fsf@gitster.siamese.dyndns.org>
In-Reply-To: <1234222843-15577-1-git-send-email-s-beyer@gmx.net>

Stephan Beyer <s-beyer@gmx.net> writes:

>   This is one of the sequencer-preparing patches.
>   (The function is used in sequencer several times, most of the time
>    with diff_flags set to DIFF_OPT_IGNORE_SUBMODULES.)
>
>   Alex is on Cc because he introduced the "Is commitable?" (i.e.
>   "Is index dirty?") part in builtin-commit.c.
>
>   Peff is on Cc because he introduced index_is_dirty() in
>   builtin-revert.c.
>
>  builtin-commit.c |   13 ++-----------
>  builtin-revert.c |   13 +------------
>  revision.c       |   15 +++++++++++++++
>  revision.h       |    2 ++
>  4 files changed, 20 insertions(+), 23 deletions(-)

It is a straightforward and clean restructuring, but please do not
contaminate revision.[ch] with this function about "internally running
diff-index".  

revision.[ch] is a library for revision/ancestry traversal and it is
already one of the largest library-ish files. It does not know nor care
about the index, and we want to keep it that way.  Please keep its focus
to revision traversal.

Perhaps diff-lib.c would be a better home for your helper function.