That's likely an oversight. It should decode based on the charset (if one exists).

Should it be fixed?

Yes. Or deprecated, as it looks like body-string is an old function and may not be relevant in modern code.