From: Junio C Hamano <gitster@pobox.com>
Date: Fri, 23 Jul 2010 21:03:49 GMT
Subject: Re: [PATCH 5/5] rebase: protect against diff.renames configuration
Message-ID: <7vvd85ewqy.fsf@alter.siamese.dyndns.org>
In-Reply-To: <AANLkTi=j8-YFwOYxcTWw3zqo=353fsB=vkmRttL4U+au@mail.gmail.com>

Sverre Rabbelier <srabbelier@gmail.com> writes:

> On Fri, Jul 23, 2010 at 12:06, Jonathan Nieder <jrnieder@gmail.com> wrote:
>> The end user configuration for "diff" should not affect the result
>> produced by the higher level command that is related to "diff" only
>> because internally it is implemented in terms of it.
>
> Almost completely unrelated and perhaps not relevant, I seem to recall
> that if you set 'ui.color' to 'always' you will get unapplyable
> patches because 'git format-patch' will include the color in it's
> output. Perhaps it should --no-color as well, while we're fixing it?

I think that it actually was a mistake for git_diff_basic_config() to call
git_color_default_config().  Parsing of diff.color.<slot> there is Ok and
was justified by 9a1805a (add a "basic" diff config callback, 2008-01-04),
but the change made to the function with 6b2f2d9 (Add color.ui variable
which globally enables colorization if set, 2008-02-18) was probably a
screw-up.  The call should instead have gone to diff_ui_config().

I also think format_config() should be changed not to call log_config()
(we need to move parsing of the format.subjectprefix from the latter to
the former), and instead call diff_basic_config() directly.