Cancellation retrieval

Problem

What becomes confusing, fragile, or impossible without understanding cancellation retrieval? This lesson answers that through explanation, a worked example, two runnable exercises, and a reference solution. No teacher-supplied worksheet is required.

Event loops let one thread coordinate many delayed operations. Promises standardized continuation and error composition over callback-heavy code.

Concept explanation

Cancellation retrieval belongs to “Async runtime rebuild”. cancellation retrieval is learned by comparing readable, constrained, and deliberately clever implementations against same evidence.

For cancellation retrieval, trace concrete input, state transition, output, and failure through a state-and-time rule spanning synchronous frames, host tasks, microtasks, rendering, cancellation, or backpressure.

Synchronous frames run to completion. Hosts schedule tasks; promise reactions enter microtask queues; rendering occurs only at host-selected opportunities. Apply that model to supplied normal, boundary, and failure cases; each case below names its input and expected evidence.

Example bank

Compare normal, boundary, failure, and cross-layer cases. Predict each observation before revealing the explanation.

Worked example

Start from supplied lab.mjs. Focus: Solve cancellation retrieval three ways using identical fixture: Build task runner with concurrency limit, AbortSignal, progress stream, and result aggregation.

1. Run: node lab.mjs
2. Save baseline evidence. Record ordering, async stacks, queue delay, cancellation signals, stream state, and cleanup paths.
3. Boundary case: Run all three implementations against: Run zero, one, and more-than-limit tasks.
4. Failure case: Force every implementation through: Reject and abort during active work, then verify queued/active cleanup.

RESULT
All versions preserve Trace never exceeds limit; result order policy is documented; all terminal paths settle and release resources.; comparison records code size, predictability, and debugging cost. Starter-level baseline: Order ends as sync:start > sync:end > microtask > task. Each line is a prefix of that final sequence.

Starter material

Create lab.mjs, paste this exact content, then run the command below.

const events = [];
const record = (value) => { events.push(value); console.log(events.join(" > ")); };

record("sync:start");
queueMicrotask(() => record("microtask:cancellation retrieval"));
setTimeout(() => record("task:cancellation retrieval"), 0);
record("sync:end");

STOP / CLEANUPScripts exit after queued work completes. Press Ctrl+C only if your experiment creates an intentional infinite loop.

Independent exercise

Compare three versions: cancellation retrieval

1. Create second case from blank file: Run all three implementations against: Run zero, one, and more-than-limit tasks.
2. Then create controlled failure: Force every implementation through: Reject and abort during active work, then verify queued/active cleanup.
3. Use Chromium Performance · Network · async debugger to prove behavior, then repair controlled failure.
4. Compare result against supplied acceptance checks and reference approach before marking complete.

Expected result

• All versions preserve Trace never exceeds limit; result order policy is documented; all terminal paths settle and release resources.; comparison records code size, predictability, and debugging cost.
• Order ends as sync:start > sync:end > microtask > task. Each line is a prefix of that final sequence.
• Controlled cancellation retrieval failure produces captured evidence; repair restores stated invariant without hiding error.

Hints

Solution

Attempt both exercises before opening reference approach.