Tomás Chor

Good idea! @navidcy did you manually cancel the tests?

> > Are there any tests for `xnodes` with stretched grids? > > There seem to be no tests for `x/y/z/λ/φnodes` whatsoever. Shall we add? I think at this point...

I'll let @glwagner comment on the possibility of an easier way, but worst case scenario, there's always the possibility of using metaprogramming to automate the creation of these functions. For...

I think if you're passing anything to `NetCDFWriter` that isn't a `Field` it can't infer dimensions from it. So you need to pass dimensions manually. [This example](https://clima.github.io/OceananigansDocumentation/stable/model_setup/output_writers/#Examples) is pretty illustrative,...

Just for reference, I created the MWE example below of a column model with a sheared `u` that evolves in time, where I write both the full `u` and an...

> Ah interesting, why do you create `coarse_model`? Is that so `coarse_model.grid` gets saved down? I think we can make it so that isn't needed probably... Right now it's needed...

> Is that a reason why you define the closure for the coarse_model? There reason is that at the moment `NetCDFWriter` needs it to get some info on the `grid`....

Created a PR in https://github.com/CliMA/Oceananigans.jl/pull/3576 to see if tests pass (i.e. if `NetCDFWriter` still works with that trivial change).

@navidcy can you please include the link to the docs preview page so that we see how it looks before approving?

> Could you build the docs locally? I know how to do it, but I didn't try. I thought the best practice was to set `push_preview=true` temporarily for docs changes...