Delete applied to private name via optional chain

[mgaudet]

I think a test case for SyntaxError on delete o?.#x needs to be added.

[rwaldron]

@leobalter does this fall in scope for the refactoring you wanted to do?

[leobalter]

This is not covered.

I did some work on optional chaining but missed covering optional chain as UnaryExpressions.

Funny enough, delete {}?.x is valid.

The version with private fields is a SyntaxError regardless of the delete.

// valid
class C {
  #x = 42;
  m() {
    var o = this;
    return o.#x;
  }
}

// SyntaxError
class C {
  #x = 42;
  m() {
    var o = this;
    return o?.#x;
  }
}

We might coverage for private fields + optional chain, but needs verification.

[mgaudet]

I don't think that second example is a syntax error since tc39/proposal-class-fields#301

[leobalter]

Thanks! I was not up to date with the private fields proposal and then we just need to verify coverage.

@mgaudet can you help me pointing to the part where delete o?.x would be forbidden?

[leobalter]

I found it: https://tc39.es/proposal-class-fields/#sec-static-semantics-early-errors

[caiolima]

IIRC, we had tests to cover that o?.#x was a SyntaxError and this got reverted into #2566. I don't remember seeing or adding any test that covers delete o?.#x though.

[rwaldron]

Is there any further action to take in Test262?