Fix the comment exclusion functionality in the lexer

[akneni]

Currently, before we can lex the source code, we clean it by removing all text in between // and \n as well as /* and */ (inclusive). We need to add some additional functionality to ignore these substrings if they're contained within a string literal.