In case a user still have old, non-UTF8 setup, they still need to see files in their encoding correctly, thus without converting output to utf-8. The way it is now, breaks things.
Also, there are cases when detected `$encoding_name` is not recognized by `decode()` subroutine.