Re: [PATCH] Teach 'git apply' to look at $GIT_DIR/config

From: Junio C Hamano <junkio@cox.net>
Date: Feb 18, 2007, 02:10 UTC
Message-ID: <7vlkiwnrl7.fsf@assigned-by-dhcp.cox.net>
In-Reply-To: <Pine.LNX.4.63.0702180301220.22628@wbgn013.biozentrum.uni-wuerzburg.de>

Johannes Schindelin <Johannes.Schindelin@gmx.de> writes:

Show 16 quoted lines

> On Sat, 17 Feb 2007, Junio C Hamano wrote:
>
>> Johannes Schindelin <Johannes.Schindelin@gmx.de> writes:
>> 
>> >> I strongly disagree from my recent day-job experience. [Explains a 
>> >> convincing use case in favour of cd-up.]
>> >
>> > Hmm. I have to think about that. But a consequence of what you suggest 
>> > would be to disallow git-apply outside of a repository, because then you 
>> > would introduce _another_ inconsistency (git-apply without --index would 
>> > behave differently when inside a repo than when outside of one).
>> 
>> Why?  git-apply outside is like 'patch' isn't it?
>
> So, why shouldn't it be the same inside? That's the inconsistency I am 
> alluding to.

It appears that you are saying we cannot make it consistent in both ways, and you are probably right.

I think git should cater first to git users who use git in git repositories, which means consistency inside a git repository between with --index and without is more important.

We can certainly teach 'git apply' without --index/--cache a new "stay where you are" option to behave like patch (but do that only without --index/--cache) inside a git repository for special cases.