for floating-point numbers [...] -convertible floating-point number
-convertible floating-point number