Reading to multiple string types in ygm::io::line_parser

[steiltre]

This PR adds a template parameter for the string type to use when reading individual lines within the ygm::io::line_parser and its derivatives.

This change allows UTF-8 files with non-ASCII characters to read to a type such as std::u32string that can appropriately handle an extended character set. This has the advantage of correctly reporting the size of strings and operator[] returning the correct characters. It has the disadvantages of requiring 4 bytes of storage for all characters (including ASCII characters that could be stored with 1 byte) and needing a user to convert the used strings before printing or writing to file (using std::codecvt or whatever replaces it since its deprecation).

Without this change, the ygm::io::line_parser is able to read all legal UTF-8 characters (from my testing using files here). Lines with non-ASCII characters read into a std::string treat the line as a collection of single-byte characters giving incorrect sizes and accessing of characters.