Currently setSignText's behavior (which may be caused by cobalt/cc-tweaked/etc.) is to convert each byte in the input to a single UTF-8 character (for example, \167 will be converted into U+00A7). This has the caveat that any unicode characters above U+00FF cannot be used. Perhaps a similar thing to SwitchCraftCC/sc-peripherals@e38b194 could be done in order to allow other unicode characters.