fix: nextTick timing regression

[mlrawlings]

Description

We recently updated the nextTick helper to run as a microtask, however the difference in timing from setImmediate has proven to cause some subtle timing issues for a few cases. We're reverting back to setImmediate-like timing using MessageChannel as a fallback which has similar timing properties and works in IE10+ postMessage because MessageChannel is not supported by JSDOM.

Because the AsyncValue was sharing this helper, but relying on it running as a microtask we now have two helpers with fallbacks: queueMicrotask and setImmediate.

Checklist:

• I have read the CONTRIBUTING document and have signed (or will sign) the CLA.
• I have updated/added documentation affected by my changes.
• I have added tests to cover my changes.