@tootle reported an issue on the mailing list: http://lists.einsteintoolkit.org/pipermail/users/2023-February/008840.html
(One of the) underlying issues is that SummationByParts does not prefix the (globally visible) Fortran subroutines with the
SBP_ prefix it uses (and must use) for globally visible symbols.
This also makes
CCTK_FNAME fail due to incorrect usage (the name it acts on must have mixed case).