Commit Graph
9 Commits
Author SHA1 Message Date
Sean McArthur f9345f99bb sync: drop old tasks in oneshot (#911) 2019-02-20 12:50:29 -08:00
Carl Lerche 3d787b16c7 sync: add loom test for mpsc (#903)
This patch updates tokio_sync::mpsc to support using loom for fuzz
testing. It includes a basic fuzz test.
2019-02-20 10:05:56 -08:00
Carl Lerche e3115231dd sync: fix mpsc/sempahore when releasing permits (#904)
This patch fixes Semaphore by adding a missing code path to the release
routine that handles the case where the waiter's node is queued in the
sempahore but has not yet been assigned the permit.

This fix is used by mpsc to handle the case when the Sender has called
`poll_ready` and is dropped before the permit is acquired.

Fixes #900
2019-02-19 16:26:05 -08:00
Sean McArthur 860ca79d62 Check Task::will_notify_current before cloning in AtomicTask 2019-02-14 13:26:53 -08:00
Stephen Carman 95b0eec8af sync: bounded channel can not have 0 size (#879) 2019-02-01 12:54:06 -08:00
Jon Gjengset 0ec8986b0b Make reason for try_send errors clearer (#864) 2019-01-23 15:06:24 -08:00
Jon Gjengset c6f8bdb249 Remove T: Debug bound on mpsc Debug impls (#866)
Following from https://github.com/tokio-rs/tokio/pull/865, this PR
removes `#[derive(Debug)]` on `mpsc` sender and receiver types in favor
of explicit `impl fmt::Debug` blocks that don't have a `T: fmt::Debug`
bound.
2019-01-23 18:04:00 -05:00
Jon Gjengset c6f9a069a5 Explicit impl Clone for tx to avoid T: Clone (#865)
`#[derive(Clone)]` on a type `struct Foo<T>` adds an impl that requires that
`T: Clone`:

```rust
impl<T: Clone> Clone for Foo<T>
```

which is unfortunate in the case of senders, because we don't want to require
that the items being sent are `Clone` for the channel sender to be `Clone`.
This PR adds an explicit `impl Clone` for the bounded and unbounded sender
types which does not have the `T: Clone` bound.

Note that this is _also_ an issue with `#[derive(Debug)]`, but that one is
harder to work around as `chan::Tx` _also_ has `#[derive(Debug)]`, as does
`chan::Chan`, so we'd have to add explicit impls for all of them to make
progress.
2019-01-23 15:51:44 -05:00
Carl Lerche 13083153aa Introduce tokio-sync crate containing synchronization primitives. (#839)
Introduce a tokio-sync crate containing useful synchronization primitives for programs
written using Tokio.

The initial release contains:

* An mpsc channel
* A oneshot channel
* A semaphore implementation
* An `AtomicTask` primitive.

The `oneshot` and `mpsc` channels are new implementations providing improved
performance characteristics. In some benchmarks, the new mpsc channel shows
up to 7x improvement over the version provided by the `futures` crate. Unfortunately,
the `oneshot` implementation only provides a slight performance improvement as it
is mostly limited by the `futures` 0.1 task system. Once updated to the `std` version
of `Future` (currently nightly only), much greater performance improvements should
be achievable by `oneshot`.

Additionally, he implementations provided here are checked using
[Loom](http://github.com/carllerche/loom/), which provides greater confidence of
correctness.
2019-01-22 11:37:26 -08:00