magiclen / boyer-moore-magiclen

Boyer-Moore-MagicLen, a fast string search algorithm implemented in Rust.

Hi, welcome to my Github pages.

Things Related to My Projects

boyer-moore-magiclen's Issues

Maybe the algo can optimize?

hello👆 git:(master) ⚡ cat src/main.rs
use std::time::Instant;
use boyer_moore_magiclen::BMByte;
/*
fn main() {
    let needle = (0..100).map(|_| "b").collect::<String>();
    let heystack = (0..100_000).map(|_| "a").collect::<String>();

    println!("Data ready.");

    let now = Instant::now();
    for _ in 0..1_000_000 {
        if heystack.find(&needle).is_some() { }
    }
    let ned = Instant::now();
    println!("cost:{:?}", ned-now);
}
*/
fn main() {
    let needle = (0..100).map(|_| "b").collect::<String>();
    let heystack = (0..100_000).map(|_| "a").collect::<String>();

    println!("Data ready.");

    let now = Instant::now();
    let bmb = BMByte::from(needle).unwrap();
    for _ in 0..1_000_000 {
        if bmb.find_first_in(&heystack).is_some() {}
    }
    let ned = Instant::now();
    println!("cost:{:?}", ned-now);
}

hello👆 git:(master) ⚡ cargo run -r   
    Finished release [optimized] target(s) in 0.01s
     Running `target/release/hello`
Data ready.
cost:4.293501137s

On my local, std::find just cost 1.381615965s

panicked at 'attempt to subtract with overflow'

Hello,
I got a crash with following code

let pattern = BMByte::from("CreateFile").expect("boyer_moore_pattern");
pattern.rfind_full_all_in("0_000_000    0008:0009:trace:rela");

text bytes [48, 95, 48, 48, 48, 95, 48, 48, 48, 32, 32, 32, 32, 48, 48, 48, 56, 58, 48, 48, 48, 57, 58, 116, 114, 97, 99, 101, 58, 114, 101, 108, 97]

stacktrace

 14:     0x562c293b8842 - boyer_moore_magiclen::byte::rfind_full::h0314f9d622769392
                               .cargo/registry/src/github.com-1ecc6299db9ec823/boyer-moore-magiclen-0.2.10/src/byte.rs:396
  15:     0x562c293b9069 - boyer_moore_magiclen::byte::BMByte::rfind_full_all_in::h05acc3dc093adc04
                               .cargo/registry/src/github.com-1ecc6299db9ec823/boyer-moore-magiclen-0.2.10/src/byte.rs:276

Add impl BMByteSearchable for Deref=[u8]

Because I don't want to create a Vec<u8> when I got a slice of [u8] or something that can be derefed to a slice of bytes like MMap.