It also compiled straightforwardly with my gfortran [gcc version 8.3.0 (Debian 8.3.0-6)] but producing a .o of 5MB.
I agree that it is not the modernized way of developing a library we’re looking for, but I guess it’s good to see previous attempts to organize the messy wold of Fortran (Open Source) libraries. Also to remind us the importance of this project and find a way out to standardize a mathematical library.
LAPACK is still written in FORTRAN 77 I believe, but Matlab, R, and Python/SciPy have nice interfaces to call it, and there is also LAPACK95, a Fortran95 interface to LAPACK. Modern Fortran interfaces could also be created for SLATEC, and I think that would take less time than recreating the functionality from scratch. OTOH, other goals of a library may include
(1) Enabling the user to understand algorithms, possibly with a view to extending them
(2) Algorithms beyond what’s in the classic FORTRAN 77 libraries.
(3) Showcasing the features of Modern Fortran