RRFS IFI crash fix + debug mode change

[SamuelTrahanNOAA]

This fixes the RRFS IFI crash, which was caused by incorrect error handling for extreme values of potential temperature.

Also, it enables libIFI's own debug mode (-DIFI_DEBUG=ON) when UPP is built with inline IFI (-I) and debug (-d). That debug mode enables bounds checking and other safeguards. It also causes uncaught C++ exceptions to pass through to Fortran. That usually results in a stack trace or core dump to the point in the C++ library that raised the exception. (Without debug mode, C++ returns an integer error code to Fortran.)

• waiting on https://github.com/NCAR/UPP_IFI/pull/5
• fixes Getting some failures with RRFS configuration #850

[WenMeng-NOAA]

@SamuelTrahanNOAA Can you sync your branch with the latest UPP develop?

[MatthewPyle-NOAA]

Great to see this problem being fixed so quickly. Can't comment on the IFI changes, but like the compile modification to utilize debug options for IFI when appropriate.

[WenMeng-NOAA]

@SamuelTrahanNOAA @MatthewPyle-NOAA
I did two tests:
test 1:
UPP version: current UPP develop
RRFS data file: RRFS parallel failure case '2023122706_case'
run directory: /u/wen.meng/ptmp/post_rrfs_2023122706-fail
** This test failed at IFI* file generation.

test 2:
UPP version: Sam's PR #852
RRFS data file: RRFS parallel failure case '2023122706_case'
run directory: /u/wen.meng/ptmp/post_rrfs_2023122706-sam-fix
** This test was successfully completed.

It appears to me that Sam's fix is working. Please let me know if you see issues in my tests.

[SamuelTrahanNOAA]

My tests finished:

• Matt Pyle's case: passed on WCOSS
• Shun Liu's case: passed on WCOSS and Hera

[WenMeng-NOAA]

@SamuelTrahanNOAA Is your PR available for final testing? If yes, please sync with the UPP develop.

[SamuelTrahanNOAA]

@WenMeng-NOAA - It is ready for testing. The libIFI PR was merged and the branch is up to date.

[WenMeng-NOAA]

The UPP RTs were completed on WCOSS2, Hera and Orion without changed results. This PR is ready for merging.