Problems

ANTLRInputStream in Swift runtime is using array of Character as internal representation to get Unicode code point as Int to supply it to the lexer. However, Swift Character is not representing Unicode code point but representing Unicode grapheme, therefore if the original String contains characters that build up from multiple Unicode code points such as Family emoji (👨‍👩‍👧‍👦, which is represented in a single Character in Swift but actually U+1F468 U+200D U+1F469 U+200D U+1F467 U+200D U+1F466), ANTLRInputStream will not able to get each Unicode code point.

Solution

• Use array of UnicodeScalar instead.
• Add unit tests to ensure ANTLRInputStream can read each unicode code point.