Character classes, whitespace and word boundaries

Understand the mechanism

A class chooses one character: [abc] does not match the entire string abc. [a-z] covers ASCII lowercase letters; [^0-9] matches one non-ASCII-digit character. Place a literal hyphen last to avoid a range.

JavaScript \d matches 0 through 9. \w primarily describes ASCII letters, digits and underscore, not all writing systems. \s includes line breaks; use [ \t] when only spaces and tabs are allowed.

The username starts with a lowercase letter, then permits lowercase letters, digits, underscore or hyphen, totaling 3 to 12 characters. \b marks a word/nonword boundary, not general natural-language segmentation.

Runnable example

const username = /^[a-z][a-z0-9_-]{2,11}$/;
console.log(JSON.stringify(["ada_7", "7ada", "ab", "a-b", "学习者"].map(s => username.test(s))));
console.log(JSON.stringify("one\t two\nthree".split(/[ \t]+/)));
console.log(JSON.stringify("cat scatter cat_2 cat!".match(/\bcat\b/g)));

Trace and verify

Username results are true, false, false, true, false. Splitting only spaces and tabs leaves the newline inside the last segment. Word matching finds the first and last cat.

Common mistakes and boundaries

Do not substitute [A-z] for [A-Za-z], because the intervening code points include punctuation. A whitelist is a business constraint, not a judgment that other names or scripts are invalid.

Practice challenge

Allow an uppercase initial while leaving subsequent characters unchanged.

Reference solution and reasoning

Change only the initial class to [A-Za-z], without an i flag. Ada passes but aDa does not, because only the first position was relaxed.

Further reading

• MDN regular expressions