Crate safina_threadpool

A threadpool.

You can use it alone or with safina, a safe async runtime.

Features

• Add closures or FnOnce to the pool
• One of the pool’s threads will execute it
• Automatically restarts panicked threads
• forbid(unsafe_code)
• Depends only on std
• 100% test coverage

Limitations

• Allocates memory
• Not optimized

Documentation

https://docs.rs/safina-threadpool

Examples

let pool =
    safina_threadpool::ThreadPool::new("worker", 2);
let receiver = {
    let (sender, receiver) =
        std::sync::mpsc::channel();
    for data in data_source {
        let sender_clone = sender.clone();
        pool.schedule(
            move || process_data(data, sender_clone));
    }
    receiver
};
let results: Vec<ProcessResult> =
    receiver.iter().collect();
// ...

Alternatives

• blocking
  • Popular
  • A little unsafe code
• threadpool
  • Popular
  • Well maintained
  • Dependencies have unsafe code
• futures-executor
  • Very popular
  • Full of unsafe
• scoped_threadpool
  • Popular
  • Contains unsafe code
• scheduled-thread-pool
  • Used by a popular connection pool library
  • Dependencies have unsafe code
• workerpool
  • Dependencies have unsafe code
• threads_pool
  • Full of unsafe
• thread-pool
  • Old
  • Dependencies have unsafe code
• tasque
  • Dependencies have unsafe code
• fast-threadpool
  • Dependencies have unsafe code
• blocking-permit
  • Full of unsafe
• rayon-core
  • Full of unsafe

Changelog

• v0.1.4 - Stop threads on drop.
• v0.1.3 - Support stable Rust! Needs 1.51+.
• v0.1.2 - Add another example
• v0.1.1 - Simplified internals and improved documentation.
• v0.1.0 - First release

TO DO

• DONE - Add schedule and try_schedule
• DONE - Add tests
• DONE - Add docs
• DONE - Publish on crates.io
• Add a stress test
• Add a benchmark. See benchmarks in https://crates.io/crates/executors
• Add a way for a job to schedule another job on the same thread, with stealing.

Release Process

1. Edit Cargo.toml and bump version number.
2. Run ./release.sh

Structs

QueueFull

Returned by try_schedule when the queue is full. This can happen when the program schedules many closures at one time. It can also happen when closures panic their threads. The pool’s throughput goes down when it must create new threads.

ThreadPool

A collection of threads and a queue for jobs (FnOnce structs) they execute.