When computing [[2,3,2][ln(2),1,5][1,8,9]] there was a memory pool
failure because of the computation of
exactAndApproximateDisplayedOutputsAreEqual. Because displaying the
right equal symbol is less important than displaying a result, we
encapsulate the method in an exception checkpoint: if there was not
enough memory on the pool to compute the equal sign, just return
EqualSign::Approximation.