Added support for [u8] inputs in Location

[danielparks]

This fixes a problem where using final_parser() on a byte string ([u8]) input would fail to compile because Location did not know how to handle anything other than str.

I tried to keep this as small and simple as possible. An alternative fix would have been to adjust Location::locate_tail() to take generic parameters that could handle both [u8] and str, but that would be significantly more complicated and harder to understand.

---

I’m happy to change things around however you want.

[Lucretiel]

Thanks for the contribution! I wasn't sure that raw bytes would usually constitute something human readable, for which line and column offsets would make sense, but I was probably being a little bit too picky.

If you could simplify it to resemble something like this:

pub fn locate_tail<'a>(original_input: &'a [u8], tail: &'a [u8]) -> Self { 
    // The original locate_tail function
}

impl RecreateContext<&[u8]> for Location {
    fn recreate_context(original_input: &[u8], tail: &[u8]) -> Self {
        Location::locate_tail original_input, tail)
    }
}

impl RecreateContext<&str> for Location {
    fn recreate_context(original_input: &str, tail: &str) -> Self {
        Location::locate_tail_bytes(original_input.as_bytes(), tail.as_bytes())
    }
}

Then I'm happy to merge