[ruby-core:120438] [Ruby master Bug#20990] Ripper.tokenize splits `"\C-\あ"` into tokens with invalid byte sequence