#uri #parser #string-parser #hyper #optimized #path

uri_parser

Simple URI parser with zero copy and easy access to all URI parts

2 unstable releases

Uses old Rust 2015

0.2.0 Oct 8, 2018
0.1.0 Nov 12, 2017

#72 in #optimized

Download history 6/week @ 2024-07-20 33/week @ 2024-07-27 32/week @ 2024-08-03 47/week @ 2024-08-10 76/week @ 2024-08-17 29/week @ 2024-08-24 2/week @ 2024-08-31 10/week @ 2024-09-07 1/week @ 2024-09-14 14/week @ 2024-09-21 7/week @ 2024-09-28 3/week @ 2024-10-05 151/week @ 2024-10-19 12/week @ 2024-10-26 295/week @ 2024-11-02

458 downloads per month

MIT/Apache

13KB
250 lines

URI parse based on nom composing parser library.

Provides simple parser that parses URI string into structure optimized for processing - Query is parsed to HashTable, path to Path etc.

Consider this code:

extern crate hyper;
extern crate uri_parser;

use uri_parser::parse_uri;
use std::time::{Duration, SystemTime};

use hyper::Uri;
use std::str::FromStr;

fn dur_f64(d: Duration) -> f64 {
    d.as_secs() as f64 + d.subsec_nanos() as f64 / 1e9
}

fn main() {
    let count = 1_000_000;
    let a_uri = "http://www.example.com/root/test?kulo=sak&kde=je&help=no&usi=yes#middle";
    let start = SystemTime::now();
    for _i in 0..count {
        let u = parse_uri(a_uri).unwrap();
        let d = u.query.unwrap();
        let h=d.get("help").unwrap();
        assert_eq!(*h, "no");
    }
    let dur = start.elapsed().unwrap();
    println!("{} loops of my parse_uri took {} secs", count, dur_f64(dur));
    let start = SystemTime::now();
    for _i in 0..count {
        let u = Uri::from_str(a_uri).unwrap();
        let q = u.query().unwrap();
        for qi in q.split("&") {
            let kv: Vec<_> = qi.split("=").collect();
            if kv[0] == "help" {
                let h = kv[1];
                assert_eq!(h, "no");
            }
        }
    }
    let dur = start.elapsed().unwrap();
    println!("{} loops of hyper from_str took {} secs", count, dur_f64(dur));
}

This library will perform better as query string is already in HashMap:

1000000 loops of my parse_uri took 0.900562534 secs
1000000 loops of hyper from_str took 1.136823832 secs

Limitations:

  1. Parses only absolute URIs
  2. Will not parse IP6 host (yet)
  3. No decoding of URL encoded strings (%hexa) - because it's referring original string
  4. Will not work well with malformed URI, only very basic parsing errors handling

Dependencies

~1MB
~22K SLoC