電子メールの検証

Validating an email "correctly" is a much harder problem than it seems: RFC 5322 is a monster. In practice we pick a good enough pattern that accepts common cases and rejects the obviously wrong ones.

Pattern: ^[\w.+-]+@[\w-]+(?:\.[\w-]+)+$

• [\w.+-]+ -- local part: letters, digits, _, ., +, -.
• @ -- separator.
• [\w-]+(?:\.[\w-]+)+ -- domain + at least one TLD separated by ..

What it accepts: mario.rossi@example.com, foo+bar@sub.example.co.uk, user_123@test-domain.io.

What it rejects: mario@, @example.com, mario@example (no TLD), spaces.

Trade-offs

The pattern above does not accept:

• Quoted strings "strange (things)"@example.com (RFC allows them).
• Domains with Unicode characters (\u00fcber@m\u00fcnchen.de).
• Very long TLDs that don't pass \w (e.g. internationalized ones).

If you need them, widen the class to \p{L} with the u flag, or delegate to a specialized library.

Trade-offs and best practices for email validation

No regex can guarantee that an email address actually exists. Overly complex patterns degrade performance and exclude unusual but valid domains. It is better to use a simple validation pattern to reject obvious typos and send a verification code.

Try it

Review exercise

Additional challenge