Update cmake build

[jedwards4b]

testing new github testing

[jedwards4b]

@billsacks Would you review this PR and also have a look at the unit tests.

[billsacks]

For now I'll document what I did and the problems I ran into. My plate is full between now and the end of June with things needed for the CESM3 science feature freeze and the upcoming ESMF release, so I won't have time before then to do more work on this - but we can circle back to this in July if you'd like to work together on it then and/or want more than a cursory review from me. (But note that I'll be gone the first and third weeks of July).

You may already know this, but the process for running these unit tests is kind of awkward - at least the process I'm aware of: the top level CMakeLists.txt file actually resides in CIME, so you need a model checkout that includes CIME and maybe some other things along with CESM_share. (See also #18 , #19 and #20 .) What I did just now was to use a CTSM checkout. Then from cime, I ran:

scripts/fortran_unit_testing/run_tests.py --build-dir `mktemp -d ./unit_test_temp.XXXXXX`

Doing this from the latest CTSM, with its out-of-the-box externals, the share tests passed.

Updating to your branch, I got these errors:

CMake Error at /Users/sacks/ctsm/ctsm0/share/src/CMakeLists.txt:23 (include):
  include could not find requested file:

    /Users/sacks/ctsm/ctsm0/cime/unit_test_temp.GTOvF8/__command_line_test__/__command_line_test__/share_src/genf90/src/genf90/CMake/genf90_utils.cmake


CMake Error at CMakeLists.txt:45 (add_subdirectory):
  add_subdirectory given source
  "/Users/sacks/ctsm/ctsm0/share/esmf_wrf_timemgr" which is not an existing
  directory.


CMake Error at CIME/non_py/src/CMake/Sourcelist_utils.cmake:105 (message):
  Source file not found: shr_kind_mod.F90

  After searching in list:
Call Stack (most recent call first):
  /Users/sacks/ctsm/ctsm0/share/test/unit/shr_abort_test/CMakeLists.txt:9 (extract_sources)


-- Configuring incomplete, errors occurred!' from dir '/Users/sacks/ctsm/ctsm0/cime/unit_test_temp.GTOvF8/__command_line_test__/__command_line_test__'

I assume this is from some moved / removed files?

Regarding the esmf_wrf_timemgr, we'd like to stop using it, but I think that doing so requires getting the value of ESMFMKFILE for the given machine/compiler in the CMake unit test build - see ESCOMP/CTSM#2375. I haven't spent much time on this, but if you know off-hand how to do this, that could help move us forward on switching over to using the actual ESMF library in the unit tests instead of esmf_wrf_timemgr. (@ekluzek @samsrabin and @adrifoster might also be interested in this.)

[jedwards4b]

If you load the ESMF module ESMFMKFILE is defined in the environment.

[jedwards4b]

 ESMF_ROOT=$NCAR_ROOT_ESMF ./scripts/fortran_unit_testing/run_tests.py --build-dir `pwd`/ftest 
Compiler is intel
Using PFUNIT_PATH: /glade/campaign/cesm/cesmdata/tools/pFUnit/pFUnit4.8.0_derecho_Intel2023.2.1_noMPI_noOpenMP
==================================================
Running cmake for __command_line_test__/__command_line_test__.
==================================================
==================================================
Running make for __command_line_test__/__command_line_test__.
==================================================
==================================================
Running CTest tests for __command_line_test__/__command_line_test__.
==================================================
Running 'ctest --output-on-failure'
Test project /glade/work/jedwards/sandboxes/cesm3_0_x/cime/ftest/__command_line_test__/__command_line_test__
      Start  1: shr_abort_mod
 1/12 Test  #1: shr_abort_mod ....................   Passed    0.29 sec
      Start  2: assert
 2/12 Test  #2: assert ...........................   Passed    0.03 sec
      Start  3: spfn
 3/12 Test  #3: spfn .............................   Passed    0.03 sec
      Start  4: infnan
 4/12 Test  #4: infnan ...........................   Passed    0.02 sec
      Start  5: shr_string
 5/12 Test  #5: shr_string .......................   Passed    0.03 sec
      Start  6: shr_strconvert
 6/12 Test  #6: shr_strconvert ...................   Passed    0.03 sec
      Start  7: shr_log
 7/12 Test  #7: shr_log ..........................   Passed    0.03 sec
      Start  8: dynamic_vector
 8/12 Test  #8: dynamic_vector ...................   Passed    0.04 sec
      Start  9: vmath
 9/12 Test  #9: vmath ............................   Passed    0.10 sec
      Start 10: shr_wv_sat_mod
10/12 Test #10: shr_wv_sat_mod ...................   Passed    0.03 sec
      Start 11: shr_precip_mod
11/12 Test #11: shr_precip_mod ...................   Passed    0.03 sec
      Start 12: shr_cal_mod
12/12 Test #12: shr_cal_mod ......................   Passed    0.28 sec

100% tests passed, 0 tests failed out of 12

Total Test time (real) =   0.95 sec

All unit tests are now working again.

[billsacks]

Awesome, thank you very much @jedwards4b !!!