This PR

• resolves Add support for computing the hypotenuse via hypot #544 by adding a specification for computing the square root of the sum of squares.
• follows IEEE 754 guidance that, if one of the operands is +-infinity, the result is infinity, regardless of whether the other operand is NaN. Intentionally, this deviates from a naive sqrt(x**2 + y**2) implementation.
• only allows underflow when both arguments are subnormal and the correct result is also subnormal. This follows C99 and the POSIX standard and effectively prohibits a naive implementation. The entire purpose of this API is to avoid underflow/overflow during intermediate steps of computation and allowing a naive implementation defeats that purpose.
• follows sqrt in requiring floating-point data types. The guidance uses "should", not "must", to allow implementations to support non-floating-point data types. The reason for the floating-point restriction is that the output data type must be a real-valued floating-point data type and promoting from, e.g., an integral to a floating-point data type is implementation-defined. For defined promotion semantics, a consumer should provide input arrays having floating-point data types.