approval-tests

Approval testing captures a known-good output as a baseline and compares future runs against it, replacing many small assertions with a single snapshot diff to reduce flaky tests and maintenance overhead.

Install the language package, write a test calling verify or verifyAsJson, run it to generate .received/.approved files, review the output, and rename the received file to approved to finalize the baseline.

Yes, golden-master testing validates large outputs, combinations, and log outputs with a single approval, supporting flexible naming and reporting across Python, Java, and JavaScript/TypeScript environments.

Approval testing supports data scrubbing to handle dynamic content and optional inline approvals, allowing you to manage expected outputs directly within the test code instead of separate files.

The workflow generates a .received file containing runtime output, compares it against the existing .approved baseline, and uses reporters to highlight diffs so you can review and update the approved reference.

Choose snapshot testing when verifying complex, large, or combination outputs where maintaining many individual assertions becomes brittle, leveraging a known-good baseline to streamline cross-language verification.