Something wrong in function

help
#1

I wanted to continue modify func in book about rust,original func:

fn parse_pair<T: FromStr>(s : &str, separator :char) -> Option<(T,T)>{
    match s.find(separator){
        None => None,
        Some(index) => {
            match (T::from_str(&s[..index]), T::from_str(&s[index+1..])){
                (Ok(l),Ok(r)) => Some((l, r)),
                _ => None
            }
        }
    }
}

My(I want to parse insted 2, three arguments) but where //1ая ветка as i suppose find the same separator,thanks)

fn parse_three<T: FromStr>(s : &str, separator :char) -> Option<(T,T,T)>{
    match s.find(separator){
        None => None,
        Some(index) => {
            match s[index+1..].find(separator){//1ая ветка
                None => match (T::from_str(&s[..index]), T::from_str(&s[index+1..])){
                    (Ok(_l),Ok(_r)) => None,  //Some((l, r,None)),
                    _ => None
                },
                Some(indexx) =>{//вторая ветка
                    match (T::from_str(&s[..index]), T::from_str(&s[index+1..indexx]),T::from_str(&s[indexx+1..])){
                        (Ok(l),Ok(r),Ok(c)) => 
                            Some((l, r,c)),
                        _ => None
                    }
                }
            }
        }
    }
}

Also question:if i want new value

let dx = (domain.1-domain.0)/steps

where dsteps is usize, and numerator isize maybe, or f32 ... Compiler has complained,but logically all is right,i want to divide on number)

#2

Here is the main problem with your code:

match s[index+1..].find(separator){//1ая ветка
// ...
  Some(indexx) =>{//вторая ветка
    match (
      T::from_str(&s[..index]), 
      T::from_str(&s[index+1..indexx]),
      T::from_str(&s[indexx+1..])) { /* ... */

indexx is the offset in &s[index+1..], not s. That is to say:

(&s[index+1..])[indexx] == s[index + 1 + indexx]

Also, this only works for characters that are 1 byte. Some characters have 2, 3, or 4 bytes.

assert!(','.len_utf8() == 1);
assert!('⚠'.len_utf8() == 3);

Here is a corrected version on the playground.

1 Like
#3

You probably want something like:

let dx: isize = (domain.1 - domain.0) as isize / steps as isize;
// or
let dx: f32 = (domain.1 - domain.0) / steps as f32;
// etc
1 Like
#4

I believe the original parse_two function is from the O'Reilly book Programming Rust, and also contains this bug. I don't see it mentioned on the errata page though. Is there someone on the forum with a copy of the book who can submit the error to O'Reilly? (Or just let me know the page number, product version, and format, and I can submit it.)

1 Like
#5

Page 42, Jim Blendy author)

1 Like
#6

I maybe wanted to use unused range in usize somehow,if maybe i will define it more than isize)) Also, do you know, this language i love very much, is it elaboration of MIT, or who is engaged in it?)