Eduard Valeyev

@loriab the CI jobs currently fail due to some branches in `int_am.cmake` not having been updated to the new component names: https://github.com/evaleev/libint/blob/e886334745e6a7e01f489e2912a0b3875a7b2dbc/cmake/modules/int_am.cmake#L129

> > > I see you bumped CMake min to 3.16, hooray! Any objections to me bumping > to 3.18 (released July 2020) to simplify int_am.cmake? > I suggest we...

> fyi that the gha file in this branch is using commas and `_LIST` options that I don't think will process as expected with cmake or int_am as originally designed....

@loriab CI passes! how painful would be to revert to cmake 3.16?

@loriab no worries BTW I am not sure I understand why we need to have the "nonlocal" (path-independent) targets. If you allow the user to discover Eigen3 on their own...

@loriab due to the header-only inefficiencies the plan to make C++ API a proper (compiled) library has been around for awhile (for as long as engine.impl.h existed). There will be...

@loriab btw I made the `int-cxx-{static,shared}` a compiled library. The header-only (interface) version is `int-cxx-headeronly-{static,shared}`. Should probably not be used by anyone since cmake is now the only way the...

This will be addressed by #205

@SpinTensor thanks for reporting this ... is the issue that `$(FCLIBS)` incomplete or that the fortran and c++ compilers are configured to use different linkers? There is no good reason...

Please provide more details. Whether `ULONGLONG_MAX` is defined checked before use: https://github.com/evaleev/libint/blob/master/include/libint2/util/intrinsic_types.h#L46 . Are you compiling generated library only, or the entire compiler+library stack? Which version (or commit)?