integrate(log(1+(1+x**2)**0.5), x) -> incorrect result



commercial macsyma gets this integral right.  I wonder if the old maxima 
got it, too,
and we broke it??