Errors

Intermediate · Functions & Errors

Go treats errors as ordinary values returned from functions, not exceptions. These exercises cover the idiomatic toolkit:

• returning and checking error values (errors.New)
• wrapping with fmt.Errorf("...: %w", err) and matching with errors.Is
• custom error types and extracting them with errors.As
• panic/recover and why it is a last resort, not flow control

Resources

• Effective Go: Errors
• Go blog: Errors are values
• Go blog: Working with Errors in Go 1.13
• Go by Example: Errors

Exercises

errors1 test

In Go, errors are ordinary values returned alongside results.

Show hint

Errors are values. Return them, do not throw them.

Import "errors" and return errors.New("...") when b == 0.
Check the result at the call site with `if err != nil`.

Source

errors2 test

Wrapping an error with %w keeps the original reachable via errors.Is.

Show hint

The %w verb wraps an error so errors.Is can unwrap and match it.

fmt.Errorf("lookup %q: %w", id, ErrNotFound)

Source

errors3 test

A custom error type is any type implementing the error interface.

Show hint

Any type with an Error() string method satisfies the error interface.

Build the message from e.Field, e.g. fmt.Sprintf("%s is required", e.Field).
errors.As walks the wrap chain to find a *ValidationError.

Source

errors4 test

panic/recover is Go’s last-resort mechanism, not normal error handling.

Show hint

recover() only works inside a deferred function.

defer func() { if r := recover(); r != nil { err = fmt.Errorf("recovered: %v", r) } }()
A named return value (err) can be set from inside the deferred func.

Source

errors5 test

errors.Join (Go 1.20) bundles several errors into one. errors.Is then

Show hint

errors.Join bundles many errors; errors.Is matches any of them:

return errors.Join(errs...)

Source