Allows construction of AnelasticDynamics and compilation of time_step! with Reactant

[dkytezab]

See #462.

[giordano]

The history looks messed up 🥲

[glwagner]

I think this is more correctly call AtmosphereModels (the logic here is that the AtmosphereModels module in the extension will define functions that pertain to src/AtmosphereModels... if that makes sense)

[giordano]

Can we remove all the unrelated commits from this PR? I can do the git surgery if you want.

[dkytezab]

that would be perfect, thank you!

[giordano]

Surgery done. Now you need to run

git fetch --all --prune
git checkout dkz/issue-build-anelastic
git reset --hard origin/dkz/issue-build-anelastic

[giordano]

Double check I didn't mess up anything 😃

[dkytezab]

LGTM, gonna just add my changes following @glwagner's suggestion

[codecov]

Codecov Report

❌ Patch coverage is 53.33333% with 7 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
ext/BreezeReactantExt/AtmosphereModels.jl	0.00%	7 Missing ⚠️

📢 Thoughts on this report? Let us know!

[dkytezab]

Same mixed topology problem: https://github.com/NumericalEarth/Breeze.jl/actions/runs/21874130501/job/63137883354?pr=463#step:7:586

[dkytezab]

See CliMA/Oceananigans.jl#5093. @giordano I wonder if we can create separate environments for tests that use Reactant and those that don't? I think we can solve this problem by bumping up the minimum version to 1.11.9 for Reactant tests

[wsmoses]

that 1.10 issue is definitely a bug that should be fixed [and from the looks of would not be limited to 1.10, we just got unlucky with julia dispatch on that version]. mwe would be helpful

[wsmoses]

1.12 should be fine [tho @giordano mentioned slow to compile], was there another reason for removing?

[wsmoses]

also any mwes for it being slow are also helpful

[glwagner]

Probably just bc of the warning that used to be emitted, right? @dkytezab put 1.12 back!

[dkytezab]

It's back!