Hi The following code does not compile with ifort 14.0.3, but compiles and works fine with gfortran(4.6, 4.7 and 4.8).
module ab_mod
type :: a
contains
procedure :: proc1
procedure :: proc2
generic :: proc => proc1, proc2
end type a
type :: b
class(a), pointer :: a
end type b
contains
subroutine proc1(this, val)
implicit none
class(a) :: this
integer, intent(out) :: val
val=1
end subroutine proc1
subroutine proc2(this, val)
implicit none
class(a) :: this
real, intent(out) :: val
val=1.0
end subroutine proc2
end module ab_mod
program test
use ab_mod
type(a), target :: a_obj
type(b) :: b_obj
integer :: val
b_obj%a => a_obj
call a_obj%proc(val)
print *, val
call b_obj%a%proc(val)
print *, val
call b_obj%a%proc1(val)
print *, val
end program test
Ifort complains that
test.f90(40): error #8486: There is no matching specific subroutine for this type bound generic subroutine call. [PROC]
call b_obj%a%proc(val)
---------------^
compilation aborted for test.f90 (code 1)
The problem centers around the naming of the class(a) component of type b. If I change type b to
type :: b
class(a), pointer :: a_obj
end type b
and make the necessary changes to the program-section, the code compiles and runs fine with ifort. The question is, is it illegal to use the original naming? If so, why does ifort not complain a line 10? And why does it complain at line 40, and not 42?
Best regards
Emil Sørensen
