use consistent argument types in EOS_Omni F90 code

[Roland Haas]

EOS_Omni was inconsistent in passing either 1-element arrays or scalars to the same Fortran subroutine. gfortran-10 by default will no longer allow this.

Pull request

https://bitbucket.org/einsteintoolkit/einsteineos/pull-requests/7/eos_omni-make-use-of-arguments-consistent

fixes this and and also fixes a (genuine) bug in the single var calls for a nuclear EOS.

[Roland Haas]

Please review.

[Roland Haas]

A second pull request handles the same type of issue in GRHydro:

https://bitbucket.org/einsteintoolkit/einsteinevolve/pull-requests/13/grhydro-use-consistent-argument-ranks-for

[Roland Haas]

Unless objected I will apply this commit after 2020-07-15

[Roland Haas]

Approved by @Steven R. Brandt

[Roland Haas]

Applied as git hash d0b8289 "EOS_Omni: account for difference between scalars and 1-element arrays" of einsteineos

[Roland Haas]

Applied as git hash ac5b9b5 "GRHydro: use consistent argument ranks for EOS_Omni calls" of einsteinevolve

[Roland Haas]

Backported to ET_2020_05 as git hash f997985 "GRHydro: use consistent argument ranks for EOS_Omni calls" of einsteinevolve