Skip to content

The Event Loop — Junior

At junior level, focus on this question:

What does the event loop actually do, in its most basic form, on every single iteration?

The core loop, in pseudocode

def event_loop():
    while True:
        ready_events = poll_os_for_ready_events(timeout=next_timer_deadline)
        for event in ready_events:
            callback = event.get_registered_callback()
            callback()  # run the code waiting for this event
        run_any_expired_timers()
flowchart LR Start["Loop starts"] --> Poll["Ask OS: 'anything\nready yet?'\n(blocks briefly if\nnothing is)"] Poll --> Ready["Get list of ready\nevents"] Ready --> Run["Run EACH ready event's\ncallback, in turn"] Run --> Start

This is the entire event loop, at its most fundamental: poll the OS for what's ready, dispatch to the corresponding registered callback for each ready item, repeat forever. Every await in your async code (per the Async/Await concurrency-model page) ultimately compiles down to "register a callback for when this operation is ready, then let the loop continue with other work."

🎓 Takeaway: no matter how sophisticated a specific async runtime's implementation gets, this poll-dispatch-repeat structure is the foundational shape every event loop shares — understanding this loop is understanding the mechanical reality underneath every await you write.

Test yourself

  1. Why does the loop need to "block briefly" during the poll step rather than spinning continuously checking for readiness?
  2. What happens to a registered callback between when it's registered and when the loop actually calls it?
  3. Why does every await ultimately reduce to "register a callback, let the loop continue"?

Continue to middle.md.