Remove invalid characters after decoding

[metapone]

In the case of an invalid array input when converting to Unicode, currently encoding.js output the invalid bytes as normal question marks (char code 63), so I can't remove invalid characters from the result without accidentally remove valid question marks. Can you ask an option to remove invalid characters after decoding, or at least use a special character instead of question mark as a placeholder for invalid result characters?

Sample code:

const targetBytes = encoding.convert(
  '?？',
  {
    to: 'SJIS',
    type: 'array',
  }
);
const result = encoding.convert(targetBytes.slice(0, 2), {
  to: 'UNICODE',
  from: 'SJIS',
  type: 'string',
});

Result: ??. Both characters are ? even though the last one should be an invalid character.