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.