JSONItem escapes ordinary Unicode characters when writing JSON

JSONItem.ToString() and JSONItem.ToFormattedString() currently emit ordinary non-ASCII string characters as Unicode escape sequences, for example Björn becomes Bj\u00F6rn. This is valid JSON, but it is surprising for users and makes generated UTF-8 JSON harder to read and compare. JSON output should preserve ordinary Unicode characters while still escaping characters that JSON syntax requires to be escaped.