FuncDefn uses wrong extension_reqs delta in inference

[acl-cqc]

1. FuncDefn stores a FunctionType (inputs, outputs, extension_reqs) as the type of the function declared. This FunctionType is what you get from the static_output of the FuncDefn node.

2. The FuncDefn node also has its own signature: FunctionType for its dataflow ports. Since it's not a dataflow node, this is:
   

   hugr/src/ops.rs

   Lines 189 to 191 in b256c2b

   	fn signature(&self) -> FunctionType {
   	Default::default()
   	}

i.e. no ports, and empty-set extension reqs.

3. In extension inference, we make the input/output extensions of the FuncDefn node equal to the extensions of the Input/Output nodes beneath it (FuncDefn is a DataflowParent just like BasicBlock, Dfg, etc.):
   

   hugr/src/extension/infer.rs

   Lines 277 to 284 in b256c2b

   	if let Some([input, output]) = hugr.get_io(node) {
   	for dir in Direction::BOTH {
   	let m_input_node = self.make_or_get_meta(input, dir);
   	self.add_constraint(m_input_node, Constraint::Equal(m_input));
   	let m_output_node = self.make_or_get_meta(output, dir);
   	self.add_constraint(m_output_node, Constraint::Equal(m_output));
   	}
   	}

4. Also in inference, we use the FunctionType from step 2 (not step 1) to constrain the input/output extensions of the FuncDefn node:
   

   hugr/src/extension/infer.rs

   Lines 305 to 315 in b256c2b

   	match node_type.signature() {
   	// Input extensions are open
   	None => {
   	let delta = node_type.op_signature().extension_reqs;
   	let c = if delta.is_empty() {
   	Constraint::Equal(m_input)
   	} else {
   	Constraint::Plus(delta, m_input)
   	};
   	self.add_constraint(m_output, c);
   	}

Putting 2,3,4 together we have that the body of the function must have a zero extension delta (Input and Output nodes of a FuncDefn are constrained to have equal extension requirements), ooops

[acl-cqc]

To be honest I find it quite confusing that you can even get a value for input_extensions or delta/extension_reqs for a FuncDefn node....