1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
//! Calloop, a Callback-based Event Loop
//!
//! This crate provides an [`EventLoop`] type, which is a small abstraction
//! over a polling system. The main difference between this crate
//! and other traditional rust event loops is that it is based on callbacks:
//! you can register several event sources, each being associated with a callback
//! closure that will be invoked whenever the associated event source generates
//! events.
//!
//! The main target use of this event loop is thus for apps that expect to spend
//! most of their time waiting for events and wishes to do so in a cheap and convenient
//! way. It is not meant for large scale high performance IO.
//!
//! ## How to use it
//!
//! For simple uses, you can just add event sources with callbacks to the event
//! loop. For example, here's a runnable program that exits after five seconds:
//!
//! ```no_run
//! use calloop::{timer::Timer, EventLoop, LoopSignal};
//!
//! fn main() {
//!     // Create the event loop. The loop is parameterised by the kind of shared
//!     // data you want the callbacks to use. In this case, we want to be able to
//!     // stop the loop when the timer fires, so we provide the loop with a
//!     // LoopSignal, which has the ability to stop the loop from within events. We
//!     // just annotate the type here; the actual data is provided later in the
//!     // run() call.
//!     let mut event_loop: EventLoop<LoopSignal> =
//!         EventLoop::try_new().expect("Failed to initialize the event loop!");
//!
//!     // Retrieve a handle. It is used to insert new sources into the event loop
//!     // It can be cloned, allowing you to insert sources from within source
//!     // callbacks.
//!     let handle = event_loop.handle();
//!
//!     // Create our event source, a timer. Note that this is also parameterised by
//!     // the data for the events it generates. We've let Rust infer that here.
//!     let source = Timer::new().expect("Failed to create timer event source!");
//!
//!     // Most event source APIs provide two things: an event source to go into the
//!     // event loop, and some way of triggering that source from elsewhere. In
//!     // this case, we use a handle to the timer to set timeouts.
//!     //
//!     // Note that this can go before or after the call to insert_source(), and
//!     // even inside another event callback.
//!     let timer_handle = source.handle();
//!     timer_handle.add_timeout(std::time::Duration::from_secs(5), "Timeout reached!");
//!
//!     // Inserting an event source takes this general form. It can also be done
//!     // from within the callback of another event source.
//!     handle
//!         .insert_source(
//!             // a type which implements the EventSource trait
//!             source,
//!             // a callback that is invoked whenever this source generates an event
//!             |event, _metadata, shared_data| {
//!                 // This callback is given 3 values:
//!                 // - the event generated by the source (in our case, a string slice)
//!                 // - &mut access to some metadata, specific to the event source (in our case, a
//!                 //   timer handle)
//!                 // - &mut access to the global shared data that was passed to EventLoop::run or
//!                 //   EventLoop::dispatch (in our case, a LoopSignal object to stop the loop)
//!                 //
//!                 // The return type is just () because nothing uses it. Some
//!                 // sources will expect a Result of some kind instead.
//!                 println!("Event fired: {}", event);
//!                 shared_data.stop();
//!             },
//!         )
//!         .expect("Failed to insert event source!");
//!
//!     // Create the shared data for our loop.
//!     let mut shared_data = event_loop.get_signal();
//!
//!     // Actually run the event loop. This will dispatch received events to their
//!     // callbacks, waiting at most 20ms for new events between each invocation of
//!     // the provided callback (pass None for the timeout argument if you want to
//!     // wait indefinitely between events).
//!     //
//!     // This is where we pass the *value* of the shared data, as a mutable
//!     // reference that will be forwarded to all your callbacks, allowing them to
//!     // share some state
//!     event_loop
//!         .run(
//!             std::time::Duration::from_millis(20),
//!             &mut shared_data,
//!             |_shared_data| {
//!                 // Finally, this is where you can insert the processing you need
//!                 // to do do between each waiting event eg. drawing logic if
//!                 // you're doing a GUI app.
//!             },
//!         )
//!         .expect("Error during event loop!");
//! }
//! ```
//!
//! ## Event source types
//!
//! The event loop is backed by an OS provided polling selector (epoll on Linux).
//!
//! This crate also provide some adapters for common event sources such as:
//!
//! - [MPSC channels](channel)
//! - [Timers](timer)
//! - [unix signals](signals) on Linux
//!
//! As well as generic objects backed by file descriptors.
//!
//! It is also possible to insert "idle" callbacks. These callbacks represent computations that
//! need to be done at some point, but are not as urgent as processing the events. These callbacks
//! are stored and then executed during [`EventLoop::dispatch`](EventLoop#method.dispatch), once all
//! events from the sources have been processed.
//!
//! ## Async/Await compatibility
//!
//! `calloop` can be used with futures, both as an executor and for monitoring Async IO.
//!
//! Activating the `executor` cargo feature will add the [`futures`] module, which provides
//! a future executor that can be inserted into an [`EventLoop`] as yet another [`EventSource`].
//!
//! IO objects can be made Async-aware via the [`LoopHandle::adapt_io`](LoopHandle#method.adapt_io)
//! method. Waking up the futures using these objects is handled by the associated [`EventLoop`]
//! directly.
//!
//! ## Custom event sources
//!
//! You can create custom event sources can will be inserted in the event loop by
//! implementing the [`EventSource`] trait. This can be done either directly from the file
//! descriptors of your source of interest, or by wrapping an other event source and further
//! processing its events. An [`EventSource`] can register more than one file descriptor and
//! aggregate them.
//!
//! ## Platforms support
//!
//! Currently, only Linux and the *BSD are supported.

#![warn(missing_docs, missing_debug_implementations)]
#![allow(clippy::needless_doctest_main)]

mod sys;

pub use sys::{Interest, Mode, Poll, Readiness, Token, TokenFactory};

pub use self::loop_logic::{EventLoop, InsertError, LoopHandle, LoopSignal, RegistrationToken};
pub use self::sources::*;

pub mod io;
mod list;
mod loop_logic;
mod sources;