zloirock/core-js · error · SyntaxError

Unexpected character after padding

Error message

Unexpected character after padding

What it means

After a complete padded chunk, the base64 spec allows no further data: once padding '=' has been consumed and remaining ASCII whitespace skipped, any non-whitespace character remaining in the string means there is data after the final padding, which is invalid. fromBase64 throws this SyntaxError rather than silently ignoring the trailing content.

Solutions

1. Remove all characters after the final '=' in the string before decoding.
2. If multiple values are being concatenated, decode each separately instead of joining the base64 strings.
3. Trim the string and confirm it ends with padding only when it is genuinely the final chunk.
4. Catch SyntaxError and reject the input at your API boundary with a clear message.

Example fix

// before
Uint8Array.fromBase64('QUJDRA==extra'); // Unexpected character after padding
// after
const idx = raw.indexOf('==');
const clean = idx !== -1 ? raw.slice(0, idx + 2) : raw;
Uint8Array.fromBase64(clean);

function hasNoTrailingData(s) {
  const idx = s.indexOf('=');
  if (idx === -1) return true;
  return /^[\t\n\f\r ]*$/.test(s.slice(idx + 1)); // only whitespace may follow first '='
}

function isBareBase64(s) {
  return typeof s === 'string' && /^[A-Za-z0-9+/=_\-\s]*$/.test(s);
}

try {
  const bytes = Uint8Array.fromBase64(input);
} catch (e) {
  if (e instanceof SyntaxError && e.message.includes('after padding')) {
    throw new TypeError('Base64 input has data after final padding — decode values separately');
  }
  throw e;
}

When it happens

Trigger: Uint8Array.fromBase64('AB==C') or any string where a '==' (or 'X=' two-byte chunk) terminated chunk is followed by more non-whitespace characters, with lastChunkHandling 'strict' or 'loose'.

Common situations: Concatenating two base64 strings where the first already ended with padding, embedding base64 inside a larger payload without delimiters, or appending suffixes (e.g. newlines mixed with stray characters, checksums) to encoded values.

Related errors

• Malformed padding: only one =
• Padding is too early
• Malformed padding: exactly one additional character
• Missing padding
• Unexpected character

AI-assisted analysis of zloirock/core-js@84e45fba09 (2026-08-30). Data as JSON: /api/errors/478f7e0d6a1f10b2. Report an issue: GitHub.

Appendix: source

Thrown at packages/core-js/internals/uint8-from-base64.js:132

    if (chr === '=') {
      if (chunk.length < 2) {
        throw new SyntaxError('Padding is too early');
      }
      index = skipAsciiWhitespace(string, index);
      if (chunk.length === 2) {
        if (index === stringLength) {
          if (lastChunkHandling === 'stop-before-partial') {
            break;
          }
          throw new SyntaxError('Malformed padding: only one =');
        }
        if (at(string, index) === '=') {
          ++index;
          index = skipAsciiWhitespace(string, index);
        }
      }
      if (index < stringLength) {
        throw new SyntaxError('Unexpected character after padding');
      }
      written = writeBytes(bytes, decodeBase64Chunk(chunk, alphabet, lastChunkHandling === 'strict'), written);
      read = stringLength;
      break;
    }
    if (!hasOwn(alphabet, chr)) {
      throw new SyntaxError('Unexpected character');
    }
    var remainingBytes = maxLength - written;
    if (remainingBytes === 1 && chunk.length === 2 || remainingBytes === 2 && chunk.length === 3) {
      // special case: we can fit exactly the number of bytes currently represented by chunk, so we were just checking for `=`
      break;
    }

    chunk += chr;
    if (chunk.length === 4) {
      written = writeBytes(bytes, decodeBase64Chunk(chunk, alphabet, false), written);
      chunk = '';

View on GitHub (pinned to 84e45fba09)