pub fn polish_tightened(r_new: Number, r_old: Number, tol: Number) -> bool
Is a polished point’s natural residual r_new tight enough, given r_old and the solver’s tol?
r_new
r_old
tol