Re: [PATCH 1/3] git remote update: Report error for non-existing groups

From: Junio C Hamano <gitster@pobox.com>
Date: Apr 8, 2009, 08:20 UTC
Message-ID: <7vy6ubo8tn.fsf@gitster.siamese.dyndns.org>
In-Reply-To: <20090408080738.GA24386@pvv.org>

Finn Arne Gangstad <finnag@pvv.org> writes:

Show 21 quoted lines

> On Tue, Apr 07, 2009 at 07:16:21PM -0700, Junio C Hamano wrote:
>> Finn Arne Gangstad <finnag@pvv.org> writes:
>> 
>> > @@ -1227,8 +1229,11 @@ static int update(int argc, const char **argv)
>> >  
>> >  	remote_group.list = &list;
>> >  	for (i = 1; i < argc; i++) {
>> > +		int groups_found = 0;
>> >  		remote_group.name = argv[i];
>> > -		result = git_config(get_remote_group, NULL);
>> > +		result = git_config(get_remote_group, &groups_found);
>> > +		if (!groups_found && (i != 1 || strcmp(argv[1], "default")))
>> > +			die("No such remote group: '%s'", argv[i]);
>> 
>> I think you are trying to be silent about the case where the caller feeds
>> you the default_argv[] array with this, but do we want to be more explicit
>> about this so that we do die when the end user explicitly says "default"
>> from the command line?
>
> Are you thinking that "git remote update default" should only be allowed
> if you have configured a group named default?

I have no preference either way, and that is why I asked.

"git remote update" without explicit "default" is obviously what your code try not to say "No such remote group" to, and that probably is a sane thing to do.

I don't know what users want to see when they say "default" explicitly without having an explicit configuration. Should it do the same thing as "git remote update"?