ast: TYPE (non-null, list and named types) does not preserve ignored characters order

[lrlna]

Description

When any of the three TYPEs (NON_NULL_TYPE, LIST_TYPE, NAMED_TYPE) contain any number of ignored character before or after the actual type, their order is not preserved, but instead gets appended to the end of the TYPE node.

Steps to reproduce

For example, the AST produced for this query,

    #[test]
    fn test_whitespace() {
        let schema = r#"
type ProductVariation {
  id: ID!,, ,
}"#;
        let parser = Parser::new(schema);

        let ast = parser.parse();
        println!("{:?}", ast);
    }

results in the following AST:

- [email protected]
    - [email protected] "\n"
    - [email protected]
        - [email protected] "type"
        - [email protected] " "
        - [email protected]
            - [email protected] "ProductVariation"
            - [email protected] " "
        - [email protected]
            - [email protected] "{"
            - [email protected] "\n  "
            - [email protected]
                - [email protected]
                    - [email protected] "id"
                - [email protected] ":"
                - [email protected] " "
                - [email protected]
                    - [email protected] ","
                    - [email protected] ","
                    - [email protected] " "
                    - [email protected] ","
                    - [email protected] "\n"
                    - [email protected]
                        - [email protected]
                            - [email protected] "ID"
            - [email protected] "}"

This can be specifically seen in just the NON_NULL_TYPE node debug output:

 - [email protected]
    - [email protected] ","
    - [email protected] ","
    - [email protected] " "
    - [email protected] ","
    - [email protected] "\n"
       - [email protected]
          - [email protected]
             - [email protected] "ID"

Expected result

The Ignored tokens should be in their correct place, and the AST for the above example's NON_NULL_TYPE node should be as follows:

 - [email protected]
       - [email protected]
          - [email protected]
             - [email protected] "ID"
    - [email protected] ","
    - [email protected] ","
    - [email protected] " "
    - [email protected] ","
    - [email protected] "\n"

This is especially useful when we implement a String representation of the AST as the goal of our library is to preserve the original input.