Modifydiag_field_add_attribute to accept r4 and r8

[uramirez8707]

Description
Modifies diag_field_add_attribute to accept r4 and r8
Adds a test

Fixes #1621

How Has This Been Tested?
CI including new test

Checklist:

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• Any dependent changes have been merged and published in downstream modules
• New check tests, if applicable, are included
• make distcheck passes

[bensonr]

This is enforcing the attribute value to be the same as the default precision specifiec at compile time, shouldn't it be the same as the incoming type-value?

[mathomp4]

I think @bensonr is right here. If I build FMS as r4 then this:

    type is (real(kind=r8_kind))
      call diag_field_add_attribute_scalar_r(diag_field_id, att_name, real(att_value))

will cast the r8 real I pass down into an r4 (since that is what real means).

In my attempt at this, I essentially "doubled up" and made two diag_field_add_attribute_scalar_r4 and a diag_field_add_attribute_scalar_r8 (and similar for the r1d → r41d, r81d)

[uramirez8707]

In the old diag manager, it ends up always writing the att_value as NF90_FLOAT.

FMS/diag_manager/diag_manager.F90

Line 4547 in d79c3b5

CALL diag_field_attribute_init(diag_field_id, att_name, NF90_FLOAT, rval=att_value)

In 4829fe8, I made the changes so that the new diag manager uses the actual precision of att_value.

[mathomp4]

If a r8 value is coming in, shouldn't this be an NF90_DOUBLE in both?

[mathomp4]

In both routines, not both lines, that is.