How can you catch a situation where server dependencies differ from those on which tests were run?

To catch situations where dependencies on the server differ from those used during testing, several approaches can be used:

• Fixing dependency versions: in package.json (npm) or yarn.lock to ensure the same versions are installed on the server.
• Automatic version checking: scripts or CI/CD pipelines that compare installed package versions with expected ones.
• Using containerization: running the application and tests in containers with identical environments.
• Local and server environment comparison: logging dependency versions at startup and comparing them.

For example, you can add a CI step that runs npm ci (clean install based on lock file) and checks for discrepancies. If dependencies differ on the server, tests may fail, and this will be immediately noticeable.