https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66409
--- Comment #5 from Steve Kargl <sgk at troutmask dot apl.washington.edu> --- On Fri, Oct 07, 2022 at 07:15:59PM +0000, anlauf at gcc dot gnu.org wrote: > > @Steve: I thought there is something in the standard that says how the > resolution (specific then generic) works, but cannot find it now... > type(*) = match any type and its type kind parameters dimension(..) = match any rank, which includes rank 0. function foo(i) integer i end foo function bar(i) type(*), dimension(..) :: i end bar Both match type "integer", kind "4", rank "0". AFAIK, there is no other consideration than TKR to discern which function to call. Perhaps, someone, who cares about this bug, can point at the language in the Fortran standard.
