阅读中文版本

Test Quality Review Workflow

Evaluate the tests in a change for meaningful assertions, edge cases, and isolation rather than mere presence.

Overview

Ensure a change ships with tests that actually verify its behaviour.

This workflow is reusable across teams and environments; adapt the commands to your tooling while keeping the sequence and checkpoints intact.

Environment and prerequisites

• A review checklist for tests
• Coverage tools
• The change under review

Procedure

1. Check assertions are meaningful

Verify tests assert observable behaviour, not just that code runs.

Expected state: Tests would fail if the behaviour regressed.

2. Check edge and error cases

Look for boundary, empty, and error-path coverage.

Expected state: Important edge cases are tested.

3. Check isolation

Confirm tests do not depend on order, shared state, or real external services.

Expected state: Tests are independent and deterministic.

4. Check the change is actually exercised

Confirm the new behaviour has a test that fails without the change.

Expected state: Every new behaviour is covered.

# confirm tests fail without the change
git stash -- src/ && npm test; git stash pop

Verification

Confirm the tests fail without the change and cover edge cases and error paths.

Decision points

Require tests for every change?

• For behaviour changes
• For every line

Recommendation: Require tests for behaviour changes; demand value, not line coverage.

Accept snapshot-only tests?

• Only with meaningful assertions too
• Snapshots suffice

Recommendation: Snapshots alone rarely catch intent regressions; pair with assertions.

Common failures

• Tests that only run code catch nothing.
• Order-dependent tests cause flakiness.
• Snapshot-only tests miss intent regressions.