Update ekat submodule

[bartgol]

Updates HAERO to use the new package-based version of ekat.

@jeff-cohere I am not 100% sure if what I did wrt to yaml/spdlog is correct; the cmake logic on that regard seemed "long". In particular,

• I no longer enable spdlog: I did not see it used in HAERO, but maybe you want it enabled b/c you want to install haero+ekat for later use?
• I always enable yaml-cpp: I removed handling of HAERO_SKIP_FIND_YAML_CPP.

In both cases, I figured you can just invoke haero's config passing EKAT_XYX=ON/OFF vars, and get the corresponding option in ekat triggered. Here, I only hard-coded ON/OF for what Haero knows for sure it's going to need...

[bartgol]

@mjschmidt271 do you guys enforce clang formatting? Here it's complaining about an empty line between headers, but I see that's done also in other files. If you want, I can address it though.

Edit: actually, I thought I did address it, with the lat commit. I honestly have no idea what clang's complaint is about... Do you know what it wants me to do? If not, then I'll remove the last commit, which is imho pointless (but I thought that was what clang complained about)

[mjschmidt271]

@bartgol We do enforce, and it looks like its primary issue is with the ordering of the ekat headers not being alphabetical. It also deletes the empty line, but that appears to be a non-necessary side effect

[bartgol]

Ok, I can make the includes alphabetical.

BTW, I don't think that's a good rule: headers should be included from most specific to most generic, to avoid hiding includes problems. E.g., if header B.hpp uses stuff from A.hpp but does not include it, it is ill-formed. But if these two are currently only used in C.hpp and included alphabetically, the error is hidden. Unfortunately, a formatter cannot check this.

That said, I will update the formatting.

[mjschmidt271]

Looks good--thank you for taking care of this!