Junio C Hamano <gitster@xxxxxxxxx> writes: > Getting rid of "config_set", which was an extra variable that kept > track of where "force" came from, does make the logic cleaner, I > guess. What we want to happen is that one of -i/-n/-f is required > when clean.requireForce is *not* unset (i.e. 0 <= require_force). Oh, noes. require_force is unset explicitly it would be 0, so this should have read (i.e. require_force != 0). Sorry for a thinko.