Fix bug in comparebfb.csh when restart data is missing in both test and base directories

[phil-blain]

If no restart files are present in neither base_dir nor test_dir, the script comparebfb.csh would simply exit with the error

foreach: No match.

with error code 1 (csh default I guess) and so the script cice.test would report

Regression baseline and test dataset are different

which is incorrect. I just added set nonomatch before the foreach so that if no restart files are present the script does not error but correctly reports :

base_data: $baseline_dir/$baseline_name/$test_name/restart/iced**
test_data: $baseline_dir/$baseline_name/$test_name/r/restart/iced**
missing data
Missing data

• Developer(s): Philippe Blain

• Please suggest code Pull Request reviewers in the column at right. @apcraig

• Are the code changes bit for bit, different at roundoff level, or more substantial? BFB

• Please include the link to test results or paste the summary block from the bottom of the testing output below.
  No tests

• Does this PR create or have dependencies on Icepack or any other models?
  No

• Is the documentation being updated with this PR? (Y/N) No
  If not, does the documentation need to be updated separately at a later time? (Y/N) No

Note: "Documentation" includes information on the wiki and .rst files in doc/source/,
which are used to create the online technical docs at https://readthedocs.org/projects/cice-consortium-cice/.

• Other Relevant Details:

I stumbled upon that bug because I was running the model for just 3 time steps with a custom test suite and had forgotten to set the dump_last namelist variable to true, which caused no restart files to be written since I was running for less than a day.

[apcraig]

Interesting. I had to google nonomatch. So, with that variable set, the foreach loop is entered with no arguments and then test_data=" "? Or is the foreach loop just skipped with no error? If the former, then I think we'll get a reasonable result (missing data). If the latter, then we won't get a reasonable result as far as I can tell. So it must be doing the former?

I think this is fine. In some ways, my preference is to check whether a valid file exists and to abort if it doesn't. But this is a reasonable workaround.

We can merge the PR.

[phil-blain]

Sorry Tony for not responding earlier, I was on vacation last week. Before this PR, the foreach loop was simply skipped with the csh error "foreach: No match.". With my modifications the foreach loop is entered, the paths are correctly formed and "missing data" is reported as expected. Thanks for merging.