naive_opt

The optimized naive string-search algorithm.

Features

The naive string-searching algorithm
Enhanced with 1-byte search like the libc++ and the libstd++ string::find
Specializing in UTF-8 strings, which is a feature of rust
The ASCII Stochastics search
Support the zero overhead trait.
minimum support: rustc 1.41.1 (f3e1a954d 2020-02-24)

Compatibility

This crate is implemented to replace the rust std library. However, the method names are different, so please rewrite your code. It shouldn't be too difficult.

compatibility:

| rust std::str | this crate | |:-----------------------------|:---------------------------------------| | std::str::find() | naive_opt::Search::search() | | std::str::rfind() | naive_opt::Search::rsearch() | | std::str::contains() | naive_opt::Search::includes() | | std::str::match_indices() | naive_opt::Search::search_indices() | | std::str::rmatch_indices() | naive_opt::Search::rsearch_indices() |

Examples

Example function:

```rust use naiveopt::{stringsearch, stringrsearch}; use naiveopt::{stringsearchindices, stringrsearchindices};

let haystack = "111 a 111b"; let needle = "a"; let r = stringsearch(haystack, needle); asserteq!(r, Some(4));

asserteq!(stringsearch(haystack, "1"), Some(0)); asserteq!(stringrsearch(haystack, "1"), Some(8));

let v: Vec<_> = stringsearchindices("abc345abc901abc", "abc").collect(); asserteq!(v, [(0, "abc"), (6, "abc"), (12, "abc")]); let v: Vec<_> = stringrsearchindices("abc345abc901abc", "abc").collect(); asserteq!(v, [(12, "abc"), (6, "abc"), (0, "abc")]); ```

Example trait: Search

```rust use naive_opt::Search;

let haystack = "111 a 111b"; let needle = "a"; let r = haystack.search(needle); assert_eq!(r, Some(4));

asserteq!(haystack.search("1"), Some(0)); asserteq!(haystack.rsearch("1"), Some(8));

let v: Vec<_> = "abc345abc901abc".searchindices("abc").collect(); asserteq!(v, [(0, "abc"), (6, "abc"), (12, "abc")]); let v: Vec<_> = "abc345abc901abc".rsearchindices("abc").collect(); asserteq!(v, [(12, "abc"), (6, "abc"), (0, "abc")]); ```

Example trait: SearchIn

```rust use naive_opt::SearchIn;

let haystack = "111 a 111b"; let needle = "a"; let r = needle.searchin(haystack); asserteq!(r, Some(4));

asserteq!("1".searchin(haystack), Some(0)); asserteq!("1".rsearchin(haystack), Some(8)); ```

Benchmark

compile by rustc 1.52.1 (9bc8c42bb 2021-05-09)

| name | bench:en | bench:ja | musl:en | musl:ja | |:------------------------|------------:|------------:|------------:|------------:| | stdstrstr | 585.580 uc | 578.320 uc | 615.830 uc | 479.560 uc | | stdstringstring | 585.960 uc | 535.340 uc | 617.810 uc | 489.860 uc | | funcstrstr | 85.994 uc | 106.520 uc | 87.228 uc | 111.430 uc | | funcstringstring | 85.223 uc | 104.990 uc | 86.863 uc | 111.340 uc | | traitstrstr | 81.330 uc | 100.390 uc | 82.727 uc | 103.500 uc | | traitstringstring | 80.777 uc | 100.920 uc | 81.788 uc | 102.520 uc | | stdindices | 527.490 uc | 402.170 uc | 514.510 uc | 394.700 uc | | funcindices | 81.626 uc | 101.900 uc | 82.144 uc | 104.220 uc | | trait_indices | 81.608 uc | 101.920 uc | 82.037 uc | 104.050 uc |

std is std::str::find()
us is micro seconds
:en is english haystack.
:ja is japanese haystack.
musl is x86_64-unknown-linux-musl
bench on intel Q6600 @ 2.40GHz

Changelogs

This crate's changelog here.