Re: [DONOTAPPLY PATCH 2/3] setup: warn about implicit worktree with $GIT_DIR

From: Jeff King <peff@peff.net>
Date: Mar 26, 2013, 20:27 UTC
Message-ID: <20130326202722.GA22769@sigill.intra.peff.net>
In-Reply-To: <20130326202142.GL1414@google.com>

On Tue, Mar 26, 2013 at 01:21:42PM -0700, Jonathan Nieder wrote:

Show 13 quoted lines

> Jeff King wrote:
> 
> > --- a/setup.c
> > +++ b/setup.c
> > @@ -437,6 +437,23 @@ const char *read_gitfile(const char *path)
> >  	return path;
> >  }
> >  
> > +static const char warn_implicit_work_tree_msg[] =
> > +N_("You have set GIT_DIR (or used --git-dir) without specifying\n"
> > +   "a working tree. In Git 2.0, the behavior will change
> 
> Please no.  I don't want git 2.0 to be delayed forever.

Please replace "2.0" with some future version, then. I just made up the number. But...

Show 5 quoted lines

> If we want this warning, would something like the following do?
> 
> 	warning: You have set GIT_DIR without setting GIT_WORK_TREE
> 	hint: In this case, GIT_WORK_TREE defaults to '.'
> 	hint: To suppress this message, set GIT_WORK_TREE='.'

That can help by teaching people how GIT_DIR behaves in general. But the warning and hint will be small consolation to somebody who runs "GIT_DIR=foo.git git clean -f" and sees it for the first time.

If you want to argue that people would see the warning in earlier runs of git, I can kind of buy that. Although the incident that triggered this discussion probably wouldn't have (I would usually start a git-clean session with "git clean" without "-f" or "git status", either of which would have done equally well as this warning to notify the user what was going on).

Like I said earlier, though, I'm not really sure this is the direction we want to go. This series is more about seeing what the fallouts are. I probably shouldn't have included this middle patch at all, because the interesting thing is what happens when we do turn it off.

-Peff