not all named sub-expressions produce their own RuleKind

[OmarTawfik]

I noticed an issue with the current parser codegen. Named sub-expressions no longer generate a RuleKind. To repro, adding ContractName to ContractDefinition below, doesn't add the new kind or change the CST. Test snapshots are unchanged as well.

NOTE: This also happens with choice: nodes, and possibly other types.

- name: "ContractDefinition"
  kind: "Rule"
  unversioned:
    sequence:
      - optional:
          terminal: "abstract"
      - terminal: "contract"
      - name: "ContractName"
        reference: "Identifier"
      - optional:
          reference: "InheritanceSpecifierList"
      - delimitedBy:
          open: "{"
          expression:
            zeroOrMore:
              reference: "ContractBodyElement"
          close: "}"

[AntonyBlakey]

Purposely doesn't work for choice:, optional:, or reference:. Not clear what it would mean in those cases.

[OmarTawfik]

Not clear what it would mean in those cases.

I would have thought they would be able to generate their own RuleKind, and insert a node with such kind in the CST.

• For reference: and choice:, this is for readability, as we might have two nodes of the same kind in succession, and we need to differentiate which is which.
• For things like optional: this is even more important, since we don't generate null/empty nodes, so in order to maintain a consistent CST shape, we need to have a named parent (with 0-1 children).