Static access shorthand parsing question

[sgrekhov]

With the Static access shorthand feature how should the following be parsed?

enum E {
  e1, e2;
  static E? one = null;
}

main() {
  var l1 = <E>[
    ....values // Error?
  ];
  var l2 = <E>[
    ... .values // ... E.values?
  ];
  var l3 = <E>[
    ...?.values // Error?
  ];
  var l4 = <E>[
    ...? .values // ...? E.values?
  ];

  var l5 = <E>[
    ?.one // Null-aware element `?E.one`?
  ];

[lrhn]

As I understand the lexical rules, tokenization is always "as far as possible".

• ....values should tokenize as ... . values which can be valid, but likely won't be unless values is a static on Iterable or Map, since that's the context type of a spread.

• ... .values. Same, and context type is Iterable<E>.

• ...?.values and ...? .values tokenize the same, context type Iterable<E>?.

• ?.one should tokenize ?. as one token which isn't valid in that position. You need ? .one for that to work. That is likely going to be an issue, I expect null-aware elements to format without a space otherwise (@munificent).