Array ops `pop_left` and `pop_right` should validate that the array type size is not 0

[doug-q]

I expect that currently we will get panics on

hugr/hugr-core/src/extension/prelude/array.rs

Line 69 in f3dafd7

array_type(TypeArg::BoundedNat { n: n - 1 }, elem_ty_var.clone());

These ops return an option type, one assumes that failure here means the array had zero size. As is this is unnecessary:

• if they array's size is TypeArg::BoundedNat{n: 0}, compute_signature should return an error (instead of panicking on the line above).
• if the array's size is a TypeArg::Variable we fail. We can't do better without arithmetic-in-the-type-system. Note this means that these pop functions are not useful inside functions polymorphic on the size of the array.

[acl-cqc]

I suspect it won't panic but n-1 will underflow back round to u64_max....

We know size can't be a TypeArg::Variable because the size is a static parameter, it has to be in order to be processed by the binary compute_signature.

So, we can drop the option and error on the n==0 case, indeed