pub fn mask_code_spans(body: &str) -> StringExpand description
_masked_code_spans: every span blanked to spaces, same length out as in.
“Same length” is Python’s, and Python counts characters: the replacement is
' ' * len(match.group()). A span holding non-ASCII therefore masks to
fewer bytes than it occupied, and a caller reasoning about columns is
reasoning about the same columns Python would.