zloirock/core-js · error · DOMException

InvalidCharacterError

InvalidCharacterError

Error message

The string contains characters outside of the Latin1 range

What it means

core-js's btoa polyfill (packages/core-js/modules/web.btoa.js:52) throws a DOMException 'InvalidCharacterError' when the input string contains a character whose char code exceeds 0xFF, because btoa can only encode Latin1 (byte) strings. This matches native window.btoa semantics: it encodes each character as one byte, so anything above U+00FF is rejected.

Source

Thrown at packages/core-js/modules/web.btoa.js:52

$({ global: true, bind: true, enumerable: true, forced: !BASIC || NO_ARG_RECEIVING_CHECK || WRONG_ARG_CONVERSION || WRONG_ARITY }, {
  btoa: function btoa(data) {
    validateArgumentsLength(arguments.length, 1);
    // `webpack` dev server bug on IE global methods - use call(fn, global, ...)
    if (BASIC) return call($btoa, globalThis, toString(data));
    var string = toString(data);
    // (string.length + 2) / 3) and then truncating to integer
    // does the ceil automatically.  << 2 will truncate the integer
    // while also doing *4.  ceil(length / 3) quanta, 4 bytes output
    // per quanta for base64.
    var output = new $Array((string.length + 2) / 3 << 2);
    var outputIndex = 0;
    var position = 0;
    var map = i2c;
    var block, charCode;
    while (charAt(string, position) || (map = '=', position % 1)) {
      charCode = charCodeAt(string, position += 3 / 4);
      if (charCode > 0xFF) {
        throw new (getBuiltIn('DOMException'))('The string contains characters outside of the Latin1 range', 'InvalidCharacterError');
      }
      block = block << 8 | charCode;
      output[outputIndex++] = charAt(map, 63 & block >> 8 - position % 1 * 8);
    } return join(output, '');
  }
});

View on GitHub (pinned to 84e45fba09)

Solutions

1. Encode to UTF-8 bytes before btoa: `btoa(String.fromCharCode(...new TextEncoder().encode(str)))` (chunk large strings to avoid argument limits).
2. Alternatively use the classic trick: `btoa(unescape(encodeURIComponent(str)))`.
3. If the environment allows, use modern `Buffer.from(str, 'utf-8').toString('base64')` (Node) or fetch-base64 helpers instead of btoa.
4. Pre-validate with a check that every char code is <= 0xFF if you intentionally require Latin1-only input.

Example fix

// before
const encoded = btoa(JSON.stringify(payload));
// after
const bytes = new TextEncoder().encode(JSON.stringify(payload));
const latin1 = Array.from(bytes, b => String.fromCharCode(b)).join('');
const encoded = btoa(latin1);

function isLatin1(s) {
  if (typeof s !== 'string') return false;
  for (let i = 0; i < s.length; i++) {
    if (s.charCodeAt(i) > 0xFF) return false;
  }
  return true;
}
if (!isLatin1(input)) {
  input = String.fromCharCode(...new TextEncoder().encode(input));
}
const encoded = btoa(input);

function isLatin1String(v) {
  return typeof v === 'string' && !/[^\u0000-\u00ff]/.test(v);
}

let encoded;
try {
  encoded = btoa(input);
} catch (e) {
  if (e.name === 'InvalidCharacterError') {
    encoded = btoa(String.fromCharCode(...new TextEncoder().encode(input)));
  } else {
    throw e;
  }
}

When it happens

Trigger: Calling `btoa(str)` where str contains any non-Latin1 character: emoji, CJK text, accented characters beyond U+00FF (e.g. 'é' is fine at U+00E9, but 'ā' U+0101, '€' U+20AC, or any emoji are not), typically from user input or decoded JSON.

Common situations: Base64-encoding user-generated content that includes Unicode; encoding JSON payloads with internationalized text; strings read from UTF-8 sources and passed directly to btoa; code that worked with ASCII test data but fails with real-world multilingual input.

Related errors

• InvalidCharacterError
• Unexpected character
• ArrayBuffer expected
• Target is not a typed array
• Object already initialized

AI-assisted analysis of zloirock/core-js@84e45fba09 (2026-08-30). Data as JSON: /api/errors/128d95e948983d4a. Report an issue: GitHub.