That's only the lexer check. We have further lints that run after parsing to detect Unicode confusables in identifiers as well, as well as two lints for text flow codepoints that can change what the code looks like in editors (one for comments and one for string literals).