Re: [wake-lock] WakeLock.request() returns a promise that never resolves (#226)

I tried to rewrite my example from https://github.com/w3c/wake-lock/issues/226#issue-477803265 with the new new new API proposal:

```js
(async () => {
  let lock = null;
  try {
    lock = await navigator.wakeLock.request('screen');
    console.log('Wake Lock is active');

    // This would fire on visibility changes
    lock.addEventListener('statechange', async (e) => {
      // Re-request a new lock after a visibility change
      lock = await navigator.wakeLock.request('screen'); // (♲)
    });

    // Schedule a regular release
    window.setTimeout(() => {
      lock.release();
      lock = null; // (←)
    }, 5000);
  } catch (e) {
    // This would catch browser interventions
    // like battery saver force-releasing the wake lock
    console.error(e.name, e.message);
    lock = null; // (←)
  }
})();
```

Two and a half questions:
- Would `onstatechange` fire in case of a browser intervention (that would be `catch`'ed (ugh, well, caught))?
- As the only state change that's possible is 'active' to 'released', maybe `onstatechange` could be renamed to `onrelease`?
- Not entirely sure about garbage collection here. If I `null`-ify the `lock` variable (above marked with ` // (←)`), would it's event listeners be GC'ed? And if I recycle the variable to re-acquire a new lock (above marked with ` // (♲)`), would a previously created event listener be still there?

-- 
GitHub Notification of comment by tomayac
Please view or discuss this issue at https://github.com/w3c/wake-lock/issues/226#issuecomment-533982302 using your GitHub account