#no-invalid-character-comparison

#Configuration

import { defineConfig, unicornPlugin } from '@rslint/core';

export default defineConfig([
  unicornPlugin.configs.recommended,
  {
    rules: {
      'unicorn/no-invalid-character-comparison': 'error',
    },
  },
]);

#Rule Details

Disallow comparing a single UTF-16 code unit from a string with a string longer than one code unit. Such equality comparisons are always false, and inequality comparisons are always true. This often catches a typo such as '/n' instead of '\n'.

The rule checks charAt(), at(), and bracket access with a static numeric key. It checks only values known to be strings, so arrays and user-defined methods are ignored. Optional character access and dynamic index keys are also ignored.

Examples of incorrect code for this rule:

'unicorn'.charAt(0) === 'ab';
'unicorn'.charAt(3) === '/n';
'unicorn'.at(0) !== 'ab';
'unicorn'[0] === 'ab';
'😀'.at(0) === '😀';

Examples of correct code for this rule:

'unicorn'.charAt(0) === 'u';
'unicorn'.charAt(3) === '\n';
'unicorn'.at(0) === 'a';
'unicorn'[0] === 'a';
'unicorn'.charAt(100) === '';

An empty string is allowed because charAt() returns it for an out-of-bounds index. Astral characters such as emoji contain two UTF-16 code units and are reported even though they appear as a single character.

When TypeScript project information is available, the rule also checks strings returned through callable types, stored in typed properties, or narrowed by checks such as typeof value === 'string'. This includes values with JSDoc types. Without project information, use a direct string annotation or assertion for values whose types cannot otherwise be determined.

#Options

This rule has no options.

#Differences from upstream

Locally declared functions or objects named String or Date are not assumed to produce strings. For example, this array comparison is allowed by rslint but reported by Unicorn v79.0.0:

function String() {
  return ['ab'];
}

String().at(0) === 'ab';

Strings containing an unpaired UTF-16 surrogate are still checked, but the quoted value in rslint's JavaScript API diagnostic message uses replacement characters. For example, for "x"[0] === "\uD800a", rslint's message quotes ���a, while Unicorn preserves the unpaired surrogate followed by a. The diagnostic range and comparison result are the same.

#Original Documentation

• eslint-plugin-unicorn: no-invalid-character-comparison
• Source code