Added backup saturation_T routine

It actually has exactly the same form as saturation_p. Interesting.  Possibility for optimization?

Signed-off-by: Ian Bell <ian.h.bell@gmail.com>

src/AbstractState.cpp

@@ -118,7 +118,7 @@ bool AbstractState::clear() {
     this->_gibbsmolar.clear();
     this->_logp.clear();
     this->_logrhomolar.clear();
-
+ 
     ///// Smoothing values
     //this->rhospline = -_HUGE;
     //this->dsplinedp = -_HUGE;