Merge main candidate 2021-02-19 #177
Conversation
Autoconf netCDF flag update
…-ncar-2020-12-15 Main candidate from NCAR (12/15/2020)
Doxygen was not assigning some docstrings for multiline variable declarations, so this patch splits those variables to one per line, with updated docstrings. Whitespace was also condensed in several of the declarations.
…update-to-GFDL-20201215 this is corresponding to GFDL main branch commit 20201218 (hash # 0f664aa)
Reduction in line length of any lines exceeding 120 characters.
Feature/update to gfdl 20201215
Fix dimensions in wave structure
- Call ice shelf initialization from within initialize_MOM.
- This removes suppport for rotation.
diagnostics and restarts during the call from initialize_MOM
- retains original drivers
Use the new scale argument to MOM_write_field to simplify write_vertgrid_file and write_ocean_geometry_file. All output and answers are bitwise identical.
Added and used the file_type as the handle for input and output, changing from using an integer. This changes the type of one of the required arguments to write_field, write_MOM_field, create_file, reopen_file, get_file_info, get_file_times, get_file_fields, write_metadata_axis and write_metadata_field, and added new variants of open_file, close_file and flush_file using this new type. Also added the new routine file_is_open. All answers are bitwise identical, but there are changes to multiple I/O interfaces.
Added the new optional scale argument to the real versions of the read_variable routines. All answers are bitwise identical, but there are new optional arguments to public routines.
Revised the depth list type used in MOM_sum_output to accommodate I/O using ordinary routines, and to make the code creating, reading and writing it potentially separable from the MOM_sum_output module by avoiding the use of the MOM_sum_output control structure in most of the routines. Also changed the index 'l' to 'li' in some places to avoid having it misread as '1'. All answers are bitwise identical but there are changes in some subroutine arguments.
Fixing an indexing error in MOM_barotropic
trailer.py PEP8/PEP257 cleanup
MOM6: +I/O interface changes, including a new file_type
- this addresses a bug in the apply_ALE_sponge routine which had been passing zero values to the remapping routine for the target grid thicknesses.
…_022321 Ale sponge fix 022321
|
@alperaltuntas, can you please reduce the line length in: Also, it looks like one of the tests in near_boundary_unit_tests is failing. I will look into that. |
|
@alperaltuntas, did you have to resolve conflicts in |
done.
Yes. Check out the following comparison vs. the original candidate. NOAA-GFDL/MOM6@dev-gfdl-main-candidate-2021-02-19...NCAR:main_candidate_2021-02-19
Yes, please. |
…021-02-19 Fix LBD module after merging main candidate 2021 02 19
|
@alperaltuntas, the unit test is still failing and I could not replicate that. Any idea why unit tests are passing in #178 and not in this PR? |
|
@gustavo-marques weird. will look into it. |
|
@gustavo-marques It looks like the MOM_lateral_boundary_diffusion unit tests are already failing in the current dev/ncar, but are passing in main_candidate_2021-02-19. I confirmed that that's the case by running the unit tests on Cheyenne with gnu. So, I believe, once this PR gets merged, the unit tests will be fixed. |
|
Ok, thanks. Will merge this now. |
Merges the main candidate from GFDL:
https://github.com/NOAA-GFDL/MOM6/pull/1328
Source code tree hasn't been changed yet, so no MOM_interface change is needed.
testing: aux_mom.cheyenne_{intel,gnu}