I don't recall the iPhone OS 3 calculator having this bug.
It seems that under some circumstances, the exponent of a floating point number overflows/underflows without triggering an error condition, or rounding to zero.
E.g. 1e-60 ^ 2 is evaluated as 1e+136.
Which is incorrect by a rather impressive 256 orders of magnitude.
It seems that under some circumstances, the exponent of a floating point number overflows/underflows without triggering an error condition, or rounding to zero.
E.g. 1e-60 ^ 2 is evaluated as 1e+136.
Which is incorrect by a rather impressive 256 orders of magnitude.