note 8 : modifying pqnl1_2 code

[LinaMiao]

-- when initial residual is less than sigma ref, correct the error of no output info
-- is the last x the optimal one? when the residual is less than sigma_ref, am I updating the x
-- minconf_pqn jumps in and out so often for FWI example, so, tighter criterion may be a better choice, not essentially needed to modify the code, try to remember does optTol can do it well