fibre 0.5.8

High-performance, safe, memory-efficient sync/async channels built for real-time, low-overhead communication in concurrent Rust applications.
Documentation 
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
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280

//! Integration tests for the SPMC (broadcast) batch send/receive APIs.

mod common;

#[allow(unused_imports)]
use common::*;

use fibre::error::{BatchSendErrorReason, RecvError, TryRecvError};
use fibre::spmc;

use std::thread;
use std::time::Duration;

// --- Sync tests ---

#[test]
fn sync_try_send_batch_all_fit() {
  let (tx, rx) = spmc::bounded::<u32>(8);
  assert_eq!(tx.try_send_batch(vec![1, 2, 3]).unwrap(), 3);
  assert_eq!(rx.try_recv_batch(8).unwrap(), vec![1, 2, 3]);
}

#[test]
fn sync_try_send_batch_partial_slowest_consumer() {
  let (tx, rx) = spmc::bounded::<u32>(4);
  let err = tx.try_send_batch((0..6).collect()).unwrap_err();
  assert_eq!(err.sent, 4);
  assert_eq!(err.reason, BatchSendErrorReason::Full);
  assert_eq!(err.unsent, vec![4, 5]);
  assert_eq!(rx.try_recv_batch(10).unwrap(), vec![0, 1, 2, 3]);
}

#[test]
fn sync_try_send_batch_no_consumers() {
  let (tx, rx) = spmc::bounded::<u32>(4);
  drop(rx);
  let err = tx.try_send_batch(vec![1, 2]).unwrap_err();
  assert_eq!(err.sent, 0);
  assert_eq!(err.reason, BatchSendErrorReason::Closed);
  assert_eq!(err.into_unsent(), vec![1, 2]);
}

#[test]
fn sync_batch_fanout_every_receiver_sees_all() {
  let (tx, rx1) = spmc::bounded::<u32>(16);
  let rx2 = rx1.clone();
  let rx3 = rx1.clone();

  assert_eq!(tx.try_send_batch((0..10).collect()).unwrap(), 10);

  for rx in [&rx1, &rx2, &rx3] {
    let mut got = Vec::new();
    while got.len() < 10 {
      rx.try_recv_batch_mut(&mut got, 4).unwrap();
    }
    assert_eq!(got, (0..10).collect::<Vec<_>>());
  }
}

#[test]
fn sync_send_batch_mut_drains_front() {
  let (tx, rx) = spmc::bounded::<u32>(3);
  let mut items = vec![10, 20, 30, 40, 50];
  assert_eq!(tx.try_send_batch_mut(&mut items).unwrap(), 3);
  assert_eq!(items, vec![40, 50]);
  assert_eq!(tx.try_send_batch_mut(&mut items).unwrap(), 0);
  assert_eq!(rx.try_recv_batch(10).unwrap(), vec![10, 20, 30]);
  assert_eq!(tx.try_send_batch_mut(&mut items).unwrap(), 2);
  assert!(items.is_empty());
}

#[test]
fn sync_send_batch_blocks_on_slow_consumer_and_unblocks() {
  let (tx, rx) = spmc::bounded::<usize>(4);

  let producer = thread::spawn(move || tx.send_batch((0..16).collect()).unwrap());

  // Give the producer time to fill the buffer and park on the slow consumer.
  thread::sleep(Duration::from_millis(100));

  let mut got = Vec::new();
  while got.len() < 16 {
    got.extend(rx.recv_batch(8).unwrap());
  }
  assert_eq!(producer.join().unwrap(), 16);
  assert_eq!(got, (0..16).collect::<Vec<_>>());
}

#[test]
fn sync_send_batch_unblocks_at_slowest_consumer_pace() {
  // One fast consumer, one slow consumer: the producer's progress is bounded
  // by the slow consumer; once the slow consumer resumes, the producer
  // completes.
  let (tx, rx_fast) = spmc::bounded::<usize>(4);
  let rx_slow = rx_fast.clone();

  let producer = thread::spawn(move || tx.send_batch((0..12).collect()).unwrap());

  let fast = thread::spawn(move || {
    let mut got = Vec::new();
    while got.len() < 12 {
      got.extend(rx_fast.recv_batch(12).unwrap());
    }
    got
  });

  // Stalled slow consumer: producer can write at most `capacity` items ahead
  // of it. Wait, then verify the producer is still blocked.
  thread::sleep(Duration::from_millis(150));
  assert!(!producer.is_finished(), "producer must block on the slowest consumer");

  // Resume the slow consumer.
  let mut slow_got = Vec::new();
  while slow_got.len() < 12 {
    slow_got.extend(rx_slow.recv_batch(4).unwrap());
  }

  assert_eq!(producer.join().unwrap(), 12);
  assert_eq!(fast.join().unwrap(), (0..12).collect::<Vec<_>>());
  assert_eq!(slow_got, (0..12).collect::<Vec<_>>());
}

#[test]
fn sync_recv_batch_blocks_until_first() {
  let (tx, rx) = spmc::bounded::<u32>(8);
  let consumer = thread::spawn(move || rx.recv_batch(8).unwrap());
  thread::sleep(Duration::from_millis(100));
  tx.try_send_batch(vec![5, 6]).unwrap();
  let got = consumer.join().unwrap();
  assert!(!got.is_empty() && got.len() <= 2);
  assert_eq!(got[0], 5);
}

#[test]
fn sync_recv_batch_disconnected_after_drain() {
  let (tx, rx) = spmc::bounded::<u32>(8);
  tx.try_send_batch(vec![1, 2, 3]).unwrap();
  drop(tx);
  assert_eq!(rx.recv_batch(10).unwrap(), vec![1, 2, 3]);
  assert_eq!(rx.recv_batch(10), Err(RecvError::Disconnected));
  assert_eq!(rx.try_recv_batch(10), Err(TryRecvError::Disconnected));
}

#[test]
fn sync_batch_wrap_around_with_drop_check() {
  use std::sync::atomic::{AtomicUsize, Ordering};
  use std::sync::Arc;

  #[derive(Clone)]
  struct Droppable(#[allow(dead_code)] u32, Arc<AtomicUsize>);
  impl Drop for Droppable {
    fn drop(&mut self) {
      self.1.fetch_add(1, Ordering::Relaxed);
    }
  }

  let drops = Arc::new(AtomicUsize::new(0));
  {
    let (tx, rx) = spmc::bounded::<Droppable>(4);
    // Multiple laps over the ring force slot overwrites (stale values must be
    // dropped in place, exactly once).
    for lap in 0..6u32 {
      let batch: Vec<Droppable> = (0..3).map(|i| Droppable(lap * 3 + i, drops.clone())).collect();
      assert_eq!(tx.try_send_batch(batch).unwrap(), 3);
      let got = rx.try_recv_batch(3).unwrap();
      assert_eq!(got.len(), 3);
      // `got` holds clones; originals stay in slots until overwritten.
    }
    drop(tx);
    drop(rx);
  }
  // 18 originals + 18 clones, all dropped exactly once.
  assert_eq!(drops.load(std::sync::atomic::Ordering::Relaxed), 36);
}

#[test]
fn sync_batch_empty_and_zero_max() {
  let (tx, rx) = spmc::bounded::<u32>(4);
  assert_eq!(tx.try_send_batch(Vec::new()).unwrap(), 0);
  assert_eq!(tx.send_batch(Vec::new()).unwrap(), 0);
  let mut empty = Vec::new();
  assert_eq!(tx.try_send_batch_mut(&mut empty).unwrap(), 0);
  assert_eq!(rx.try_recv_batch(0).unwrap(), Vec::<u32>::new());
  assert_eq!(rx.recv_batch(0).unwrap(), Vec::<u32>::new());
}

// --- Async tests ---

#[cfg(not(miri))]
mod async_tests {
  use super::*;
  use tokio::time::timeout;

  #[tokio::test]
  async fn async_send_batch_all_fit() {
    let (tx, rx) = spmc::bounded_async::<u32>(8);
    assert_eq!(tx.send_batch(vec![1, 2, 3]).await.unwrap(), 3);
    assert_eq!(rx.recv_batch(8).await.unwrap(), vec![1, 2, 3]);
  }

  #[tokio::test]
  async fn async_send_batch_blocks_then_completes() {
    let (tx, rx) = spmc::bounded_async::<usize>(4);
    let send_task = tokio::spawn(async move { tx.send_batch((0..16).collect()).await.unwrap() });

    let mut got = Vec::new();
    while got.len() < 16 {
      let batch = timeout(LONG_TIMEOUT, rx.recv_batch(8)).await.unwrap().unwrap();
      got.extend(batch);
    }
    assert_eq!(send_task.await.unwrap(), 16);
    assert_eq!(got, (0..16).collect::<Vec<_>>());
  }

  #[tokio::test]
  async fn async_send_batch_mut_cancel_safe() {
    let (tx, rx) = spmc::bounded_async::<u32>(2);
    let mut items = vec![1, 2, 3, 4, 5];
    {
      let fut = tx.send_batch_mut(&mut items);
      let res = timeout(Duration::from_millis(100), fut).await;
      assert!(res.is_err(), "future should be pending: buffer fills at 2");
    }
    assert_eq!(items, vec![3, 4, 5]);
    assert_eq!(rx.try_recv_batch(10).unwrap(), vec![1, 2]);
  }

  #[tokio::test]
  async fn async_send_batch_closed_mid_batch() {
    let (tx, rx) = spmc::bounded_async::<u32>(2);
    let send_task = tokio::spawn(async move { tx.send_batch((0..10).collect()).await });

    tokio::time::sleep(Duration::from_millis(50)).await;
    drop(rx);

    let err = send_task.await.unwrap().unwrap_err();
    assert_eq!(err.sent + err.unsent.len(), 10);
    assert_eq!(err.unsent, ((err.sent as u32)..10).collect::<Vec<_>>());
  }

  #[tokio::test]
  async fn async_fanout_recv_batch() {
    let (tx, rx1) = spmc::bounded_async::<u32>(16);
    let rx2 = rx1.clone();

    assert_eq!(tx.send_batch((0..8).collect()).await.unwrap(), 8);

    for rx in [rx1, rx2] {
      let mut got = Vec::new();
      while got.len() < 8 {
        let mut out = Vec::new();
        let k = rx.recv_batch_mut(&mut out, 3).await.unwrap();
        assert_eq!(k, out.len());
        got.extend(out);
      }
      assert_eq!(got, (0..8).collect::<Vec<_>>());
    }
  }

  #[tokio::test]
  async fn async_recv_batch_cancel_safe() {
    let (tx, rx) = spmc::bounded_async::<u32>(4);
    {
      let fut = rx.recv_batch(4);
      let res = timeout(Duration::from_millis(50), fut).await;
      assert!(res.is_err(), "future should be pending on empty channel");
    }
    tx.try_send_batch(vec![1, 2]).unwrap();
    assert_eq!(rx.recv_batch(4).await.unwrap(), vec![1, 2]);
  }

  #[tokio::test]
  async fn async_recv_batch_disconnected() {
    let (tx, rx) = spmc::bounded_async::<u32>(4);
    tx.try_send_batch(vec![7]).unwrap();
    drop(tx);
    assert_eq!(rx.recv_batch(4).await.unwrap(), vec![7]);
    assert_eq!(rx.recv_batch(4).await, Err(RecvError::Disconnected));
  }
}