Re: [PATCH 5/5] builtin-prune: protect objects listed on the command line

From: Michele Ballabio <barra_cuda@katamail.com>
Date: Mar 27, 2008, 21:11 UTC
Message-ID: <200803272211.44126.barra_cuda@katamail.com>
In-Reply-To: <alpine.LSU.1.00.0803271735010.3802@racer.site>

On Thursday 27 March 2008, Johannes Schindelin wrote:

Show 19 quoted lines

> On Thu, 27 Mar 2008, Junio C Hamano wrote:
> 
> > Junio C Hamano <junio@pobox.com> writes:
> > 
> > > From: Junio C Hamano <gitster@pobox.com>
> > > Date: Mon, 24 Mar 2008 23:20:51 -0700
> > >
> > > Finally, this resurrects the documented behaviour to protect other 
> > > objects listed on the command line from getting pruned.
> > >
> > > Signed-off-by: Junio C Hamano <gitster@pobox.com> ---
> > >  * This is done deliberately differently from what you did.  Because 
> > >    we do not want to accept "we allow losing what's reachable from 
> > >    master" with "git prune master..next", setup_revisions() is not the 
> > >    right thing to use for this command.
> > 
> > Ping?
> 
> I did not see any problem with your implementation

Me neither, but, as a nitpick, wouldn't something like

		if (!get_sha1(name, sha1)) {
			object = parse_object(sha1);
			if (!object)
				die("bad object %s", name);
		} else
			die("unrecognized argument: %s", name);

be a bit more useful?