React

React

Purpose

Write React where state lives in one place, effects are rare, and re-renders are understood rather than suppressed with memoization applied at random.

When to Use

• Building or reviewing React components.
• Diagnosing unnecessary re-renders or stale state.
• Deciding where state belongs: local, lifted, context, or server.
• Removing useEffect calls that should not exist.

Capabilities

• Component decomposition and state colocation.
• Hook correctness: dependencies, cleanup, and the rules that are not optional.
• State management selection: useState, useReducer, context, external store, server state.
• Data fetching with TanStack Query or the framework's own loader.
• Render profiling and targeted memoization.

Inputs

• The component tree and where data enters it.
• The interaction and its performance characteristics, if performance is the concern.
• React version — the correct answer changed with 18 and again with 19.

Outputs

• Components with a single source of truth for each piece of state.
• Effects only where genuinely synchronizing with an external system.
• Memoization applied where a profile shows it is needed.

Workflow

1. Locate the state — Put it as close to where it is used as possible. Lift only when two siblings need it. Reach for context only when prop drilling exceeds about three levels.
2. Separate server state from client state — Data from an API is cache, not state. It has staleness, refetching, and error semantics that useState does not model. Use a query library.
3. Delete unnecessary effects — An effect that computes a value from props belongs in render. An effect that resets state on a prop change belongs in a key. Most useEffect calls in a typical codebase should not exist.
4. Profile before memoizing — React DevTools Profiler shows what actually re-renders and why. useMemo on a cheap computation costs more than it saves.
5. Make the dependencies honest — Never silence the exhaustive-deps lint rule. If the array is wrong, the bug is a stale closure, and it will be intermittent.

Best Practices

• Derived state is a bug. If a value can be computed from props or other state, compute it during render.
• useEffect is for synchronizing with something outside React: a subscription, a DOM API, a timer. It is not for reacting to state changes.
• Every effect that subscribes must return a cleanup function. Missing cleanup is the standard cause of memory leaks and duplicate listeners.
• Do not put a non-stable key on a list. Index keys break every time the list is reordered or filtered.
• Context re-renders every consumer when its value changes. Split contexts by update frequency, or use an external store with selectors.
• Lift state up only as far as needed. State in the root component re-renders the tree.

Examples

An effect that should not exist:

// Wrong: derived state, an extra render, and a chance to be out of sync.
function Cart({ items }) {
  const [total, setTotal] = useState(0);
  useEffect(() => {
    setTotal(items.reduce((sum, i) => sum + i.price * i.qty, 0));
  }, [items]);
  return <Total value={total} />;
}

// Right: compute it during render. It is always correct, by construction.
function Cart({ items }) {
  const total = items.reduce((sum, i) => sum + i.price * i.qty, 0);
  return <Total value={total} />;
}

Server state belongs in a query, not in useState plus useEffect:

function OrderList({ status }) {
  const { data, isPending, error } = useQuery({
    queryKey: ["orders", status],
    queryFn: ({ signal }) => fetchOrders(status, { signal }),
    staleTime: 30_000,
  });

  if (isPending) return <Skeleton />;
  if (error) return <ErrorState error={error} onRetry={() => refetch()} />;
  return <List items={data} />;
}

The manual version needs loading state, error state, cancellation on unmount, a race-condition guard when status changes mid-flight, and a cache. That is what the library is.

Notes

• The React Compiler (React 19) auto-memoizes and removes most hand-written useMemo and useCallback. Do not spend effort on memoization you are about to delete.
• Strict Mode in development intentionally double-invokes effects to surface missing cleanup. An effect that breaks under Strict Mode is broken in production too — it just fails less often.
• key on a component is the idiomatic way to reset its state when an identity changes. It is far cleaner than an effect that resets fields.