Review brief

Beginner20 min

An identity service is centralizing username normalization so account creation, sign-in, and uniqueness checks use one stable representation. The caller supplies arbitrary user-entered Unicode text, and the returned value becomes the stored uniqueness key.

Behavioral contract

Normalize with Unicode NFKC first, trim leading and trailing whitespace second, and apply locale-independent lowercase third.
The canonical value must contain only ASCII lowercase letters, digits, and underscores; hyphens and every other character are invalid and must throw RangeError rather than be rewritten.
Validate the canonical value at 3 through 20 Unicode code points inclusive.
Equivalent inputs must produce the same canonical value on every host locale because the result is the stored uniqueness key.

Review instructions

1Review the changed function against the ordered canonicalization contract.
2Anchor each finding to the narrowest useful changed line and explain the identity impact.
3Recommend a direction rather than writing replacement code.
4Prioritize no more than four findings.

Learning objectives

• Apply normalization steps in a specified order
• Validate the canonical value rather than its source spelling
• Recognize locale-dependent identity behavior

src/username-normalization.tsTypeScript

@@ -0,0 +1,13 @@

1+export function normalizeUsername(input: string): string {

2+ const candidate = input.trim().toLocaleLowerCase();

4+ if (input.length < 3 || input.length > 20) {

5+ throw new RangeError("Username must be between 3 and 20 characters");

6+ }

8+ if (!/^[a-z0-9_-]+$/.test(candidate)) {

9+ throw new RangeError("Username contains unsupported characters");

10+ }

11+

12+ return candidate.replaceAll("-", "_");

13+}

Reference solution

The findings below are ordered by priority.

Findings

1. Unicode compatibility forms never reach the canonical representation

Location: src/username-normalization.ts:1-2

The function trims and lowercases the original spelling without applying NFKC. Compatibility-equivalent input such as full-width Latin characters is rejected instead of resolving to the same ASCII uniqueness key as its canonical spelling.

Apply NFKC before trimming and lowercasing, then validate the resulting value.

2. Canonical length is validated against the source spelling

Location: src/username-normalization.ts:4-6

The length check reads input.length, before trimming or compatibility normalization, and counts UTF-16 code units rather than Unicode code points. Surrounding whitespace can reject an otherwise valid username, while normalization can change the length that is actually stored.

After all normalization steps, count code points in the canonical value and enforce the inclusive 3-to-20 range there.

3. Hyphens are silently converted into underscores

Location: src/username-normalization.ts:8-12

The validation explicitly accepts hyphens and then rewrites them. The contract requires rejection, and the rewrite makes distinct-looking submissions such as team-one and team_one compete for the same key without telling the caller why.

Allow only letters, digits, and underscores in the canonical value and throw for a hyphen.

4. Lowercasing depends on the process locale

Location: src/username-normalization.ts:2

Calling toLocaleLowerCase without a fixed locale lets host configuration affect identity. The same input can therefore normalize differently across service instances, undermining uniqueness and sign-in lookup.

Use the specified locale-independent lowercase operation after NFKC.

Reasonable non-findings

Trimming before validation is required; only its position after NFKC matters.
A RangeError is the documented failure type for invalid canonical values.
Returning one string is appropriate because persistence and collision reporting belong to the caller.