eslint-plugin-dot-notation-unicode

English | 日本語

An ESLint plugin that extends the dot-notation rule with full Unicode identifier support. Fixes the issue where ESLint's built-in rule doesn't recognize Japanese, Chinese, Korean, and other Unicode characters as valid identifiers.

The Problem

ESLint's built-in dot-notation rule uses this regex for identifier validation:

const validIdentifier = /^[a-zA-Z_$][a-zA-Z0-9_$]*$/u;

This only matches ASCII characters, so valid JavaScript like obj.日本語 is not recognized, and obj["日本語"] won't be flagged for conversion.

The Solution

This plugin uses Unicode property escapes to correctly identify valid ECMAScript identifiers:

const validIdentifier = /^[\p{ID_Start}_$][\p{ID_Continue}$\u200C\u200D]*$/u;

Installation

npm install --save-dev eslint-plugin-dot-notation-unicode
# or
yarn add -D eslint-plugin-dot-notation-unicode

Usage

Flat Config (ESLint 9+)

// eslint.config.js
import dotNotationUnicode from "eslint-plugin-dot-notation-unicode";

export default [
  {
    plugins: {
      "dot-notation-unicode": dotNotationUnicode,
    },
    rules: {
      "dot-notation": "off",
      "dot-notation-unicode/dot-notation": "error",
    },
  },
];

Legacy Config

{
  "plugins": ["dot-notation-unicode"],
  "rules": {
    "dot-notation": "off",
    "dot-notation-unicode/dot-notation": "error"
  }
}

Examples

Invalid (will be auto-fixed)

obj["日本語"]    // → obj.日本語
obj["한국어"]    // → obj.한국어
obj["中文"]      // → obj.中文
obj["français"]  // → obj.français
obj["العربية"]   // → obj.العربية
obj["english"]   // → obj.english

Valid (no error)

obj.日本語
obj.한국어
obj.中文
obj.français
obj.العربية
obj.english
obj["content-type"]  // hyphen is not valid in identifiers
obj["123start"]      // cannot start with a number
obj[variable]        // dynamic access

Options

This plugin supports the same options as ESLint's built-in dot-notation rule, plus an additional option for legacy parser compatibility.

allowKeywords (default: true)

Set to false to follow ECMAScript version 3 compatible style, requiring bracket notation for reserved word properties.

// With { "allowKeywords": false }
obj["class"]  // OK
obj.class     // Error

allowPattern (default: "")

Allows bracket notation for property names that match the specified regular expression pattern. This is useful when working with external APIs that use snake_case while maintaining camelCase in your own code.

// With { "allowPattern": "^[a-z]+_[a-z]+$" }
obj["snake_case"]  // OK (matches pattern)
obj["camelCase"]   // Error (doesn't match pattern)

legacyParserSupport (default: false)

Set to true for compatibility with older parsers (TypeScript < 5.5, Node.js < 18.20 / 19.x / < 20.12). When enabled, certain Unicode characters that are valid in the latest Unicode specification but not recognized by older parsers will not be converted to dot notation.

Affected characters:

• ・ (U+30FB) KATAKANA MIDDLE DOT
• ･ (U+FF65) HALFWIDTH KATAKANA MIDDLE DOT
• Zero Width Non-Joiner (U+200C)
• Zero Width Joiner (U+200D)

// With { "legacyParserSupport": true }
obj["send・receive"]  // OK (not converted, "・" may cause issues in older parsers)

Note: This option is only needed if you're using TypeScript < 5.5 or Node.js < 18.20 / 19.x / < 20.12. These characters were fixed in Unicode 15.1.

Why a Separate Plugin?

• ESLint's dot-notation rule is frozen and not accepting new features
• @typescript-eslint/dot-notation inherits the same limitation and requires type information (slower)
• No existing npm package solves this issue

References

• ESLint dot-notation rule
• Unicode UAX #31 (Identifier Specification)
• ECMAScript IdentifierName

License

MIT License - see LICENSE for details.