After spending some time trying to understand how KIND works in Fortran I'm still in doubt. I would like to know if, e.g., INTEGER(KIND = c_intptr_t) is equivalent toINTEGER(c_intptr_t)?
If yes, which way is the preferred way?
As a type declaration the two forms (for kind parameter k) integer (kind = k) and integer (k) are indeed equivalent.
This is true in type declaration statements
integer (kind=k) x
integer (k) y ! x and y are integers with kind parameter k
and in other places
x = [integer(kind=k) :: 1, 2, 3]
y = [integer(k) :: 1, 2, 3]
The same holds for the intrinsic types real(kind=k), complex(kind=k) and logical(kind=k). One has to be careful with character:
character(kind = k) c ! Length 1, kind k
character(k) d ! Length k, default kind
character(l, k) e ! Length l, kind k
Preferred form is subjective and you should follow your style guide consistently.
For derived types this is not necessarily the case:
type t(k)
integer, kind :: k=58
end type t
type(t(kind=12)) :: x ! Not valid: kind parameter is not "kind"
type(t(k=12)) :: y
type(t(12)) :: z
end