Bug in atparse script

[zhanglikate]

What is wrong?

It seems there's a mistake in the atparse script. The issue lies in how the while loop is structured within the atparse function. Specifically, the loop termination condition [[ "$__after" == "$__text" ]] may not always work as expected, leading to premature termination of the loop and missing the last line. So the last line in the file of diag_table_aod is always missing in the finale diag_table, which results in the missing of sea salt AOD in the diagnostic output.

Steps to reproduce

In the run directory, you can see that the last line of diag_table is always missing compared to the diag_table_template

Additional information

"gfs_phys", "DU_AOD_550", "du_aod550", "fv3_history2d", "all", .false., "none", 2
"gfs_phys", "SU_AOD_550", "su_aod550", "fv3_history2d", "all", .false., "none", 2
"gfs_phys", "BC_AOD_550", "bc_aod550", "fv3_history2d", "all", .false., "none", 2
"gfs_phys", "OC_AOD_550", "oc_aod550", "fv3_history2d", "all", .false., "none", 2
"gfs_phys", "SS_AOD_550", "ss_aod550", "fv3_history2d", "all", .false., "none", 2
The last line is always missing in the diag_table

Do you have a proposed solution?
In the atparse script, I've changed the while loop condition to while IFS= read -r __text || [ -n "$__text" ], which ensures that the loop continues until the input is fully consumed, even if the last line doesn't end with a newline character. It works well to solve the issue.

[SamuelTrahanNOAA]

Your fix would have spawned a subprocess. The atparse.bash is carefully written not to spawn subprocesses. I have a slightly different fix here:

• regression test system bug fixes, eliminate MOM6 warnings (#2197), add xr_cnvcld flag to FV3 (#2185) #2202

Plus, it has lots of comments and an updated test program.