https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86006
--- Comment #3 from Dominique d'Humieres <dominiq at lps dot ens.fr> --- > From my understanding the compiler translates the generic call > at compile time into calls for the functions bound to the generic call. > If this is the case, and given the fact that the compiler distinguishes > explicitly between dummy arguments of type "integer" and of type > "unlimited polymorphic pointer", there should be no ambiguity. AFAICT looking at examples, intrinsics types and kinds are part of unlimited polymorphic. I had a quick look at the standard and saw some restrictions for allocatable and pointer, but nothing saying that they can be used to resolve ambiguities. > I'll trigger a discussion at the intel ifort forum and let you know. Could you please give a pointer to the thread? You could also do it on comp.lang.fortran.