Lexer is panicking with some weird unicode chars

[bnjjj]

Here is a test to reproduce:

#[test]
fn tests_unicode() {
    let gql_1 = "\"\n\n\\u{c}\nPSK\\u{1}\\0\\0\\0י";
    let lexer_1 = Lexer::new(gql_1);
    dbg!(lexer_1.tokens);
    dbg!(lexer_1.errors);
}

it will panic because it doesn't respect char boundary here

[lrlna]

It panics because one of the (line terminator) characters in that particular string doesn't get added to the Error data, so we end up with an off by one error, which means we can't slice the original input string properly.

As a quick workaround, you may wish to put this string into a raw string: r#"\"\n\n\\u{c}\nPSK\\u{1}\\0\\0\\0י"#. I'll be looking into this bug tomorrow.