Is this really a “modernized” version of SLATEC? It doesn’t appear as such, meaning the TYPE, RANK, and INTERFACE mismatches have never been permitted by the standard - as explicitly as shown any way.
And Fortran compilers at least since the days of Fortran 95 i.e., circa 1997 and thereafter have generally issued diagnostics toward this.
g95 must have been rather permissive to let such issues pass.
Perhaps SLATEC functionality (along with modernized refactoring of existing code if licensing would allow) can be “rolled” into the current “standard library” effort and coded in a standard-conforming manner?