Should CustomType bounds be derivable from the type args?

[aborgna-q]

Some Custom types may be able to specify stricter bounds for some type instantiations.
For example, a List[T] may be Bound.Any in general but Bound.Copyable if T: Copyable.

Note that that cannot be assumed in general. We could for example define a custom type resembling a tuple (qubit, T), in which case the type would always be Bound.Any independently of T.

Should we have some mechanism to compute the bounds given on instantiation?

[aborgna-q]

Nvm, this is already implemented for typedefs

hugr/hugr-core/src/extension/type_def.rs

Lines 14 to 23 in abd70c9

	/// The type bound of a [`TypeDef`]
	#[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
	#[serde(tag = "b")]
	#[allow(missing_docs)]
	pub enum TypeDefBound {
	/// Defined by an explicit bound.
	Explicit { bound: TypeBound },
	/// Derived as the least upper bound of the marked parameters.
	FromParams { indices: Vec<usize> },
	}

(it's just missing from the spec)