On Wed, Jul 23, 2025 at 05:53:17PM +0100, Paul Richard Thomas wrote:
>
> The attached implements the F2018 generic statement, which has the same
> semantics as the typebound version but can appear in any specification
> statement.
>
> As it says in the first comment in the patch, use is made of the existing,
> typebound matching functions to obtain access-spec and generic-spec. After
> this the standard INTERFACE machinery is used.
>
> I spent a stupidly long time allowing the mixing of generic statements with
> generic interfaces until I realised that I was accepting ST_GENERIC in the
> wrong place in parse_spec :-(
>
> Regtests on x86_64/FC42 - OK for mainline?
>
Hi Paul,
I think the code following my .sig should compile and execute
% gfcx -o z a.F90 && ./z
42.0000000
43.5000000
% gfcx -o z -DGEN a.F90
a.F90:9:21:
9 | generic :: bar => bar, bah
| 1
Error: There's already a non-generic procedure with binding
name 'bar' at (1)
The 'interface bar ... end interface bar' is an idiom
that I use throughout my math library.
--
steve
module foo
implicit none
private
public bar
#ifdef GEN
generic :: bar => bar, bah
#else
!
! Should be equivalent to above.
!
interface bar
module procedure bar
module procedure bah
end interface bar
#endif
contains
function bar(i)
real bar
integer, intent(in) :: i
bar = i
end function bar
function bah(x)
real bah
real, intent(in) :: x
bah = x
end function bah
end module foo
program snooze
use foo
print *, bar(42)
print *, bar(43.5)
end program snooze
