Add support for passing integer(8) values to mpas_log_write.

[jim-p-w]

This PR allows callers of mpas_log_write to provide values of type integer(8).

The placeholder specification is $w for integer(8) values, as opposed to $i for integer values.

I ran a test where I interspersed integer and integer(8) values, e.g.

integer, dimension(:), allocatable :: intarray
integer(8), dimension(:), allocatable :: int8array
...
call mpas_log_write('SMIOLf_get_var error $i i81:$w int1:$i i82:$w int2:$i i83:$w', intArgs=[local_ierr, intarray(1), intarray(2)], int8Args=int8array, messageType=MPAS_LOG_ERR)

[mgduda]

@jim-p-w Can you add more detail to the commit message?

[jim-p-w]

@jim-p-w Can you add more detail to the commit message?

@mgduda How about:
Allow passing values of type integer(8)/integer(kind=I8KIND) to mpas_log_write.
The string $j is used to identify integer(kind=I8KIND) values.

I'm open to suggestions about what the commit message says.

[abishekg7]

@mgduda Just for my understanding, would a core infrastructure addition like this require an addition to the unit tests also?

[abishekg7]

@jim-p-w Probably a basic question, but I'm trying this code snippet out in your branch,

 integer (kind=I8KIND) :: tmp_i8_var = 5123123_I8KIND

 call mpas_log_write('rk_setup i8 kind log $w', int8Args=(/tmp_i8_var/))

and I get this as output. Do I need to set some kind of formatting width here, or am I doing something else wrong perhaps?

 rk_setup i8 kind log **

[jim-p-w]

@jim-p-w Probably a basic question, but I'm trying this code snippet out in your branch,

 integer (kind=I8KIND) :: tmp_i8_var = 5123123_I8KIND

 call mpas_log_write('rk_setup i8 kind log $w', int8Args=(/tmp_i8_var/))

and I get this as output. Do I need to set some kind of formatting width here, or am I doing something else wrong perhaps?

 rk_setup i8 kind log **

@abishekg7 I changed the placeholder string to $j instead of $w per Michaels suggestion.

[jim-p-w]

@mgduda Just for my understanding, would a core infrastructure addition like this require an addition to the unit tests also?

@abishekg7 @mgduda It wasn't clear to me how to create a unit test for mpas_log_write. When I look at the existing tests in src/core_test I see mpas_log_write is assumed to work, as it is used all over the test code.

[abishekg7]

@mgduda Just for my understanding, would a core infrastructure addition like this require an addition to the unit tests also?

@abishekg7 @mgduda It wasn't clear to me how to create a unit test for mpas_log_write. When I look at the existing tests in src/core_test I see mpas_log_write is assumed to work, as it is used all over the test code.

Yeah, I also don't see any specific tests for mpas_log_write itself.

Otherwise, the PR looks good. I've tested with NVHPC GPU and CPU versions so far.