From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Fri, 30 Jan 2009 13:23:02 GMT
Subject: Re: [PATCH] Switch receive.denyCurrentBranch to "refuse"
Message-ID: <alpine.DEB.1.00.0901301421170.3586@pacific.mpi-cbg.de>
In-Reply-To: <20090130073415.GA27224@coredump.intra.peff.net>

Hi,

On Fri, 30 Jan 2009, Jeff King wrote:

> On Fri, Jan 30, 2009 at 08:17:48AM +0100, Johannes Sixt wrote:
> 
> > Johannes Schindelin schrieb:
> > > +		error("refusing to update checked out branch: %s\n"
> > > +			"if you know what you are doing, you can allow it by "
> > > +			"setting\n\n"
> > > +			"\tgit config receive.denyCurrentBranch true\n", name);
> > 
> > Oh, fscking hell, I should have screamed out loudly when Jeff named this
> > option "denyCurrentBranch" instead of "allowCurrentBranch". It's all too
> > easy to fall into the trap, like you here.
> 
> Sorry. ;P
> 
> On the other hand, you also missed the boat on receive.denyDeletes and
> receive.denyNonFastForwards.

The idea with these is that they are _booleans_, and therefore

	[receive]
		denyDeletes

is something natural to write, because "denyDeletes" is _not_ the default.

However, with denyCurrentBranch we wanted to change the default in the 
long run, so I agree it was a not-so-brilliant choice.

Ciao,
Dscho