Comment by paulddraper

This is just tracing GC. Nothing more or less than that.

The only difference is how you construct the closure record for your GC.

Using the example:

  function f() {
    const a = {};
    const b = {};
    return [
      () => a,
      () => b,
    ];
  }

You can choose to have one scope record shared by both functions, or individual record.

  function1 ----|   ----> a
                v   |
                scope
                ^   |
  function2 ----|   ----> b

Or

  function1 --> scope1 --> a

  function2 --> scope2 --> b

JS engines choose to do the former, not out of simplicity (arguably it's more complicated), but due to performance.