str::splitn(N, ' ') is slower than it ought to be

[jeffsmith82]

I wrote a small benchmark and was surprised at how much slower it was using str::splitn rather than using bytes directly.

splitn
Time elapsed in parse_response() is: 2.290853462s

as bytes
Time elapsed in parse_response() is: 661.947666ms

Is it possible to do something similar to #46693 where if splitting on a single byte.

https://users.rust-lang.org/t/performance-comparison/56041/2

use std::time::{Duration, Instant};
fn main() {
    let response = String::from("HTTP/1.1 418 I'm a teapot\r\n");
    let mut res: (&str, &str, &str) = ("", "", "");
    let start = Instant::now();
    for _ in 0..100_000_000 {
        res = match parse_http(&response) {
            Ok(data) => data,
            Err(_) => {
                continue;
            }
        };
    }
    let duration = start.elapsed();

    println!("version:{}\ncode:{}\ndescription:{}\n", res.0, res.1, res.2);
    println!("Time elapsed in parse_response() is: {:?}", duration);
    
    let start2 = Instant::now();
    for _ in 0..100_000_000 {
        res = match parse_http2(&response) {
            Ok(data) => data,
            Err(_) => {
                continue;
            }
        };
    }
    let duration2 = start2.elapsed();

    println!("version:{}\ncode:{}\ndescription:{}\n", res.0, res.1, res.2);
    println!("Time elapsed in parse_response() is: {:?}", duration2);
}

fn parse_http(s: &str) -> Result<(&str, &str, &str), &str> {
    let mut parts = s.splitn(3, ' ');
    let version = parts.next().ok_or("No Version")?;
    let code = parts.next().ok_or("No status code")?;
    let description = parts.next().ok_or("No description")?;
    Ok((version, code, description))
}

fn parse_http2(s: &str) -> Result<(&str, &str, &str), &str> {
    let mut bytes = s.as_bytes().iter();
    let i = bytes.position(|b| *b == b' ').ok_or("No Version")?;
    let j = bytes.position(|b| *b == b' ').ok_or("No status code")? + i + 1;

    let version = &s[..i];
    let code = &s[(i + 1)..j];
    let description = &s[(j + 1)..];
    Ok((version, code, description))
}

[bikbov]

Hi. I would like to take this issue.

[mati865]

@rustbot assign @bikbov

You can use @[DELETE_THIS]rustbot claim (obviously, without [DETLETE_THIS]) to assign yourself.

[rustbot]

Error: Parsing assign command in comment failed: ...'ign bikbov' | error: user should start with @ at >| '

You can '...

Please file an issue on GitHub at triagebot if there's a problem with this bot, or reach out on #t-infra on Zulip.

[mati865]

@rustbot assign @bikbov

[bikbov]

@rustbot claim

[bikbov]

@rustbot assign @bikbov

You can use @[DELETE_THIS]rustbot claim (obviously, without [DETLETE_THIS]) to assign yourself.

thank you

[bend-n]

its worth mentioning that the changes made in #46735 (at least in 2025) already modify this code path; large gains cant really be expected, at least for this extremely small string.

(and besides, memchr isnt really that good for extremely small strings, and rustc's memchr is not that good at all.)