The implementation of pow is incorrect
pow(0.0 + 0_e, 0.0) returns 1 + -nan_e. In this case I expect that the return value should be 1 + 0_e.
pow(0.0 + 0_e, 0.0) returns 1 + -nan_e. In this case I expect that the return value should be 1 + 0_e.