Use worksize for folded-grid extents

[taimoorsohail]

Summary

• replace four suspicious size(grid) extent calculations with worksize(grid)
• align folded-grid setup paths with RightFaceFolded launch semantics
• keep the change scoped to the pressure, ERA5, bathymetry, and tracer-diffusion paths we identified

Why

On RightFaceFolded grids, worksize(grid) includes the extra folded northern row while size(grid) does not. These setup paths could therefore build auxiliary fields over 1:Ny while runtime kernels later operate over the folded 1:Ny+1 working domain, which is a plausible source of seam-local instability.

Impact

This keeps extent calculations consistent with folded-grid kernel launches and should reduce the risk of mismatched north-fold state near the seam.

Validation

• inspected the diff to confirm the change is limited to the four intended sites
• did not run local tests, per repository instructions

[codecov]

Codecov Report

❌ Patch coverage is 0% with 28 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/Atmospheres/hydrostatic_pressure.jl	0.00%	26 Missing ⚠️
src/Bathymetry/regrid_bathymetry.jl	0.00%	1 Missing ⚠️
src/InitialConditions/diffuse_tracers.jl	0.00%	1 Missing ⚠️

📢 Thoughts on this report? Let us know!

[navidcy]

this seems like something for a kernel launch?

[simone-silvestri]

yeah probably this should be a kernel, I think claude really doesn't like kernels

[taimoorsohail]

This loop is in main, not introduced by this PR, so I think this is an existing bug (not defending Claude of course ;) )

[taimoorsohail]

Have a look now @simone-silvestri @navidcy

[simone-silvestri]

looks good

[navidcy]

The CI breaks.

Was wondering, do we want to change the Base.size definition?
But then Base.size gives the number of cells, right? So perhaps we want to have both size and worksize?