Stage 4#411Coding · Senior~10 min read

Test operation merge logic exhaustively

Testing + Tooling Drills · Unit + Integration Tests

Cover local-only pending, backend-only records, duplicate tx hashes, stale entries, and final reconciliation.

Prompt

This is a hands-on rep. Attempt the drill before reading the model answer, then narrate the tradeoffs as if an interviewer is watching.

💡

Tip

This is a better interview signal than shallow component snapshots.

Foundation

What to ground before answering

Cover local-only pending, backend-only records, duplicate tx hashes, stale entries, and final reconciliation.

Focus vocabulary: Vitest, merge logic, edge cases.

The useful mental model is not to memorize a perfect answer. It is to explain what owns the data, what can fail, what the user sees, and what test would prove the behavior.

Coding drill

Choose high-value merge test cases

Target: 15mNot run

Make the behavior executable before comparing against the model answer.

type MergeRisk = 'local-only' | 'backend-only' | 'duplicate-tx' | 'stale-local' | 'reconciled';

function mergeTestCaseNames(risks: MergeRisk[]): string[] {
  // TODO: return stable, human-readable test names for each merge risk.
  return [];
}

TypeScript · runnable

System design

Interview explanation prompt

What problem is this practice item really testing?
What state or contract boundary must be explicit?
What edge case would cause a production regression?
What would you test first?
How would you explain the tradeoff in two minutes?

Self-grade

Strong answer handles the edge cases before polishing syntax.
Strong answer explains why the chosen type or function boundary prevents bugs.
Weak answer passes only the happy path or hides uncertainty in booleans and nullable fields.

Model Answer

function mergeTestCaseNames(risks: MergeRisk[]): string[] {
  const names: Record<MergeRisk, string> = {
    'local-only': 'keeps local pending operation visible',
    'backend-only': 'shows backend transfer without local record',
    'duplicate-tx': 'deduplicates local and backend records by tx hash',
    'stale-local': 'expires stale local pending operation',
    reconciled: 'clears local pending after reconciliation',
  };
  return risks.map((risk) => names[risk]);
}

Review

Recall before moving on

What is the one-sentence answer for "Test operation merge logic exhaustively"?
Which real experience from PR TIMES, React/TypeScript migration, or systems work supports it?
What edge case would you volunteer before the interviewer asks?
What is the smallest test or artifact that proves the design works?