Re: [PATCH 2/2] mkfs: remove notion of config "type"

[Dave Chinner <david@fromorbit.com>]

On Thu, Jun 14, 2018 at 09:41:21AM -0700, Darrick J. Wong wrote:
> On Wed, Jun 13, 2018 at 02:37:01PM -0500, Eric Sandeen wrote:
> > @@ -3796,19 +3795,18 @@ _("respecification of configuration not allowed\n"));
> >  		ret = parse_defaults_file(fd, &dft, config_file);
> >  		if (ret) {
> >  			fprintf(stderr,
> > -_("Error parsing %s config file: %s : %s\n"),
> > -				default_type_str(dft.type),
> > +_("Error parsing config file: %s : %s\n"),
> >  				config_file, strerror(errno));
> >  			free(config_file);
> >  			close(fd);
> >  			exit(1);
> >  		}
> > +		printf(_("Configuration file used for defaults: %s\n"),
> 
> _("EXPERIMENTAL configuration file used for defaults: %s\n")

I don't see a reason for this to be experimental - we're not
screwing with on-disk formats and there's no potential for data
corruption or loss, so why do we need to mark this as "don't use
this because it might screw up your data"?  I'd much prefer we spend
the time now to get it right before before shipping it...

Which then makes me ask: why don't we have a config file version
identifier in the config file?

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com