Support builds with all ACCESS types

[dougiesquire]

Also removed the generic tracers git submodule, which will no longer be needed

[aidanheerdegen]

Suggest to add a comment, up to you.

[dougiesquire]

Note, please do not merge this. This compiles, but the legacy ACCESS-OM-BGC type segfaults when run. So I still have some work to do here.

[aidanheerdegen]

Note, please do not merge this

I have converted to draft.

[dougiesquire]

Okay, I've tested the changes in this PR in our different models use prereleases and the repro testing CI:

Model	Configuration tested	Historical reproducibility
ACCESS-OM2	dev-1deg_jra55_ryf dougiesquire/1deg_jra55_ryf_wombatlite	❌ Test PR here ❌ Test PR here
ACCESS-OM2-BGC (legacy)	dev-1deg_jra55_ryf_bgc	✅ Test PR here
ACCESS-ESM1.6	dev-preindustrial+concentrations	✅ Test PR here

The answer changes to the ACCESS-OM2+WOMBATlite configuration are only in WOMBAT tracers and are due to updates to WOMBATlite since the config was last updated.

Similarly, I think the answer changes to the ACCESS-OM2 configuration come from previous changes on the MOM5 development branch, not the changes in this PR (namely, the changes in this PR).

So I believe this is ready to merge.

(Note, these tests required changes to the MOM5 SPR. I used this branch, but @harshula is working on a better implementation of the functionality in that branch.)

[dougiesquire]

@aidanheerdegen would you mind taking another quick look?

This is what has changed since you last looked: I added back in the allocation of some arrays so that we can build with legacy WOMBAT if we want to (using the ACCESS-OM-BGC build type)

[aidanheerdegen]

Looks fine to me, though I must admit to being a bit confused by all the simultaneous changes ...

[dougiesquire]

Looks fine to me, though I must admit to being a bit confused by all the simultaneous changes ...

Yeah, there are a lot of dependent changes taking place