function-missing-result (S221)

What it does

Checks that every function explicitly declares which variable contains the result.

Why is this bad?

Specifying the result() clause allows for the variable inside the function containing the result to be named differently from the function.

Allowing the variable to be named different from the function allows for better naming of both the function and the internal variables, and also it can help when creating functions that might need to be duplicated to have versions supporting various types. Since declaring the result in the result() clause allows for easier duplication of the function because then less lines must be changed.

Example

interface distance
    module procedure distance_real32, distance_real64
end interface distance

function distance_real32(a, b)
    real(kind=real32), intent(in) :: a, b
    real(kind=real32)             :: distance_real32

    distance_real32 = abs(a-b)
end function distance_real

function distance_real64(a, b)
    real(kind=real64), intent(in) :: a, b
    real(kind=real64)             :: distance_real64

    distance_real64 = abs(a-b)
end function distance_real

Use instead:

interface distance
    module procedure distance_real32, distance_real64
end interface distance

function distance_real32(a, b) result(distance)
    integer, parameter               :: func_type = real32

    real(kind=func_type), intent(in) :: a, b
    real(kind=func_type)             :: distance

    distance = abs(a-b)
end function distance_real

function distance_real64(a, b) result(distance)
    integer, parameter               :: func_type = real64

    real(kind=func_type), intent(in) :: a, b
    real(kind=func_type)             :: distance

    distance = abs(a-b)
end function distance_real

References

• Clerman, N. Spector, W., 2012, Modern Fortran: Style and Usage, Cambridge University Press, Rule 129, p. 178.