Fix the CreateDiagTest function

[jgfouca]

It was dropping additional arguments

Fixes #7554

[BFB]

[jgfouca]

Dang, it looks like the np>1 tests do not work!

[mahf708]

You mean they do NOT pass, but they do work, right?

And yep, that's precisely why we need those tests, because some of these diags are rank sensitive and can easily break ;)

The following tests FAILED:
	230 - zonal_avg_np2 (Failed)
	231 - zonal_avg_np3 (Failed)
	232 - zonal_avg_np4 (Failed)
CMake Error at /home/runner/_work/E3SM/E3SM/components/eamxx/cmake/ctest_script.cmake:76 (message):
  Test had fails

[jgfouca]

@mahf708 , correct. They run but do not pass. I have added @cjvogl

[mahf708]

If you prefer, we can disable the np>1 tests for zonal_avg and make an issue to reenable them later. They never worked before. I do intend to introduce np>1 tests for other diagnostics (e.g., #7551). I can also take a look into the zonal_avg fails later (no rush to fix them right away)

[mahf708]

Actually, it shouldn't be a hard fix. Should I push? I think the scaled area inside zonal avg needs a comm-wide call for a sum to rescale it. (See horiz_avg for an example of how to do that)

[jgfouca]

@mahf708 , yes, plz push a fix if you have one.

[mahf708]

@mahf708 , yes, plz push a fix if you have one.

I take it back... I don't have an immediate fix, but I will look harder...

[cjvogl]

The "now uncovered" zonal average issue with np>1 is in the unit test itself (and thankfully not the diagnostic). I can migrate the fix from #7481 here or I'm happy with the approach of disabling the zonal average np>1 checks in this PR (and I'll re-enable them in #7481).

[jgfouca]

@cjvogl , I don't want to have to remember to re-enable something, so if you can migrate the fix (assuming it's small) that would be great.

[jgfouca]

Thanks, @cjvogl !

[cjvogl]

Thanks, @cjvogl !

Thanks for catching the CreateDiagTest bug!