What is the recommended practice for handling exceptions in library code to avoid breaking clients?

• A. Map to stable domain errors, document error types, avoid leaking internals, and provide consistent error information.
• B. Throw raw internal exceptions to expose details.
• C. Return only generic strings with no error codes.
• D. Always crash the application on any error.

Answer: (A)

When a library communicates errors, the surface it exposes to clients needs to be stable and predictable. The best approach is to map failures to stable domain errors, clearly document the error types, avoid leaking internal implementation details, and provide consistent information such as error codes or structured messages. This means clients can rely on a known set of errors and handle them deterministically, even as the library’s internals evolve. By keeping the error surface stable and well-documented, you allow users to write robust error-handling logic, improve debuggability, and prevent breaking clients when internal changes occur.

Throwing raw internal exceptions leaks implementation details and creates brittle APIs: clients might depend on internal types or messages that can change, causing failures to propagate in unexpected ways. Returning only generic strings with no error codes deprives clients of a machine-readable way to distinguish error kinds or act on them programmatically. Always crashing the application on any error is user-unfriendly and unacceptable for a reusable library, which should fail gracefully and offer recoverable paths.