The float arg checks now check against all the numpy float types (float16, float32, float64, float128).
Authored by: bugman 2012-02-15
Parent: [r15334]
Child: [r15336]