This is my substring search workspace.

Documentation

Fast substring search for strings and byte strings, using the two-way algorithm.

This is the same code as is included in Rust's libstd to “power” str::find(&str), but here it is exposed with some improvements:

Available for byte string searches using &[u8]
Having an optional SSE4.2 accelerated version (if detected at runtime) which is even faster.
Using memchr for the single byte case, which is ultra fast.
twoway::find_bytes(text: &[u8], pattern: &[u8]) -> Option<usize>
twoway::rfind_bytes(text: &[u8], pattern: &[u8]) -> Option<usize>
twoway::find_str(text: &str, pattern: &str) -> Option<usize>
twoway::rfind_str(text: &str, pattern: &str) -> Option<usize>

Recent Changes

0.2.1
- Update dev-deps
0.2.0
- Use std::arch and transparently support SSE4.2 when possible (x86 and x86-64 only) to enable an accelerated implementation of the algorithm. Forward search only. By @RReverser and @bluss
- Fix a bug in the SSE4.2 algorithm that made it much slower than it should have been, so performance increases as well.
- Requires Rust 1.27
0.1.8
- Tweak crate keywords by @tari
- Only testing and benchmarking changes otherwise (no changes to the crate itself)
0.1.7
- The crate is optionally no_std. Regular and pcmp both support this mode.
0.1.6
- The hidden and internal test module set, technically pub, was removed from standard compilation.
0.1.5
- Update from an odds dependency to using unchecked-index instead (only used by the pcmp feature).
- The hidden and internal test module tw, technically pub, was removed from standard compilation.
0.1.4
- Update memchr dependency to 2.0
0.1.3
- Link to docs.rs docs
- Drop pcmp's itertools dependency
- Update nightly code for recent changes
0.1.2
- Internal improvements to the pcmp module.
0.1.1
- Add rfind_bytes, rfind_str
0.1.0
- Initial release
- Add find_bytes, find_str

MIT / APACHE-2.0