This is what engineering told me:
You can actually detect if a character is an emoji icon yourself in Lua, but it’ll take quite a bit of work.
The key thing to remember here is that a string that gets pushed into Lua is UTF-8 encoded, which can be made up of both ASCII and Unicode characters. This means that a single character can range between 1 and 4 bytes in your string. This is why UTF-8 strings are sometimes referred to as multibytes strings.
You can identify a character in your string as a unicode character by checking its highest bit (ie: the 8th bit). If the last bit is set, then the preceding bits indicate how many bytes represent that 1 unicode character. This is documented here…
http://en.wikipedia.org/wiki/UTF-8#Description
You can access the individual bytes of the string by using the following function in Lua…
http://docs.coronalabs.com/api/library/string/byte.html
From there, you have to extract the unicode value yourself. Once you’ve done that, you can recognize if it is an emoji icon by checking if it’s unicode character is within the range specified here…
http://en.wikipedia.org/wiki/Emoji#Regional_Indicator_Symbols
But that said, if you’re trying to restrict the string for something like a login user name and password, then it’s simpler to restrict the characters to English characters and maybe western European character codes.