vortex-io 0.69.0

Core async and blocking IO traits and implementations, used by IPC and file format
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
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
// SPDX-License-Identifier: Apache-2.0
// SPDX-FileCopyrightText: Copyright the Vortex contributors

use std::sync::Arc;

use futures::Stream;
use futures::StreamExt;
use futures::stream::BoxStream;
use smol::block_on;

use crate::runtime::BlockingRuntime;
use crate::runtime::Executor;
use crate::runtime::Handle;
pub use crate::runtime::pool::CurrentThreadWorkerPool;

/// A current thread runtime allows callers to much more explicitly drive Vortex futures than with
/// a Tokio runtime.
///
/// The current thread runtime will do no work unless `block_on` is called. In other words, the
/// default behavior is single-threaded with code running on the thread that called `block_on`.
///
/// It's also possible to clone the runtime onto other threads, each of which can call `block_on`
/// to drive work on that thread. Each thread shares the same underlying executor with the same
/// set of tasks, allowing work to be driven in parallel.
///
/// For automatic driving of work, a [`CurrentThreadWorkerPool`] can be created from the runtime
/// by calling [`new_pool`](CurrentThreadRuntime::new_pool). The returned pool can be configured
/// with the desired number of worker threads that will drive work on behalf of the runtime.
#[derive(Clone, Default)]
pub struct CurrentThreadRuntime {
    executor: Arc<smol::Executor<'static>>,
}

impl CurrentThreadRuntime {
    /// Create a new current thread runtime.
    pub fn new() -> Self {
        Self::default()
    }

    /// Create a new worker pool for driving the runtime in the background.
    ///
    /// This pool can be used to offload work from the current thread to a set of worker threads
    /// that will drive the runtime's executor.
    ///
    /// By default, the pool has no worker threads; the caller must set the desired number of
    /// worker threads using the `set_workers` method on the returned pool.
    pub fn new_pool(&self) -> CurrentThreadWorkerPool {
        CurrentThreadWorkerPool::new(Arc::clone(&self.executor))
    }

    /// Returns an iterator wrapper around a stream, blocking the current thread for each item.
    ///
    /// ## Multi-threaded Usage
    ///
    /// To drive the iterator from multiple threads, simply clone it and call `next()` on each
    /// clone. Results on each thread are ordered with respect to the stream, but there is no
    /// ordering guarantee between threads.
    pub fn block_on_stream_thread_safe<F, S, R>(&self, f: F) -> ThreadSafeIterator<R>
    where
        F: FnOnce(Handle) -> S,
        S: Stream<Item = R> + Send + 'static,
        R: Send + 'static,
    {
        let stream = f(self.handle());

        // We create an MPMC result channel and spawn a task to drive the stream and send results.
        // This allows multiple worker threads to drive the execution while all waiting for results
        // on the channel.
        let (result_tx, result_rx) = kanal::bounded_async(1);
        self.executor
            .spawn(async move {
                futures::pin_mut!(stream);
                while let Some(item) = stream.next().await {
                    // If all receivers are dropped, we stop driving the stream.
                    if let Err(e) = result_tx.send(item).await {
                        tracing::trace!("all receivers dropped, stopping stream: {}", e);
                        break;
                    }
                }
            })
            .detach();

        ThreadSafeIterator {
            executor: Arc::clone(&self.executor),
            results: result_rx,
        }
    }
}

impl BlockingRuntime for CurrentThreadRuntime {
    type BlockingIterator<'a, R: 'a> = CurrentThreadIterator<'a, R>;

    fn handle(&self) -> Handle {
        let executor: Arc<dyn Executor> = Arc::clone(&self.executor) as Arc<dyn Executor>;
        Handle::new(Arc::downgrade(&executor))
    }

    fn block_on<Fut, R>(&self, fut: Fut) -> R
    where
        Fut: Future<Output = R>,
    {
        block_on(self.executor.run(fut))
    }

    fn block_on_stream<'a, S, R>(&self, stream: S) -> Self::BlockingIterator<'a, R>
    where
        S: Stream<Item = R> + Send + 'a,
        R: Send + 'a,
    {
        CurrentThreadIterator {
            executor: Arc::clone(&self.executor),
            stream: stream.boxed(),
        }
    }
}

/// An iterator that wraps up a stream to drive it using the current thread execution.
pub struct CurrentThreadIterator<'a, T> {
    executor: Arc<smol::Executor<'static>>,
    stream: BoxStream<'a, T>,
}

impl<T> Iterator for CurrentThreadIterator<'_, T> {
    type Item = T;

    fn next(&mut self) -> Option<Self::Item> {
        block_on(self.executor.run(self.stream.next()))
    }
}

/// An iterator that drives a stream from multiple threads.
pub struct ThreadSafeIterator<T> {
    executor: Arc<smol::Executor<'static>>,
    results: kanal::AsyncReceiver<T>,
}

// Manual clone implementation since `T` does not need to be `Clone`.
impl<T> Clone for ThreadSafeIterator<T> {
    fn clone(&self) -> Self {
        Self {
            executor: Arc::clone(&self.executor),
            results: self.results.clone(),
        }
    }
}

impl<T> Iterator for ThreadSafeIterator<T> {
    type Item = T;

    fn next(&mut self) -> Option<Self::Item> {
        block_on(self.executor.run(self.results.recv())).ok()
    }
}

#[expect(clippy::if_then_some_else_none)] // Clippy is wrong when if/else has await.
#[cfg(test)]
mod tests {
    use std::sync::Arc;
    use std::sync::Barrier;
    use std::sync::atomic::AtomicUsize;
    use std::sync::atomic::Ordering;
    use std::thread;
    use std::time::Duration;

    use futures::StreamExt;
    use futures::stream;
    use parking_lot::Mutex;

    use super::*;

    #[test]
    fn test_worker_thread() {
        let runtime = CurrentThreadRuntime::new();

        // We spawn a future that sets a value on a separate thread.
        let value = Arc::new(AtomicUsize::new(0));
        let value2 = Arc::clone(&value);
        runtime
            .handle()
            .spawn(async move {
                value2.store(42, Ordering::SeqCst);
            })
            .detach();

        // By default, nothing has driven the executor, so the value should still be 0.
        assert_eq!(value.load(Ordering::SeqCst), 0);

        // An empty pool still does nothing.
        let pool = runtime.new_pool();
        assert_eq!(value.load(Ordering::SeqCst), 0);

        // Adding a worker thread should drive the executor.
        pool.set_workers(1);
        for _ in 0..10 {
            if value.load(Ordering::SeqCst) == 42 {
                break;
            }
            thread::sleep(Duration::from_millis(10));
        }
        assert_eq!(value.load(Ordering::SeqCst), 42);
    }

    #[test]
    fn test_block_on_stream_single_thread() {
        let mut iter =
            CurrentThreadRuntime::new().block_on_stream(stream::iter(vec![1, 2, 3, 4, 5]).boxed());

        assert_eq!(iter.next(), Some(1));
        assert_eq!(iter.next(), Some(2));
        assert_eq!(iter.next(), Some(3));
        assert_eq!(iter.next(), Some(4));
        assert_eq!(iter.next(), Some(5));
        assert_eq!(iter.next(), None);
    }

    #[test]
    fn test_block_on_stream_multiple_threads() {
        let counter = Arc::new(AtomicUsize::new(0));
        let num_threads = 4;
        let items_per_thread = 25;
        let total_items = 100;

        let iter = CurrentThreadRuntime::new()
            .block_on_stream_thread_safe(|_h| stream::iter(0..total_items).boxed());

        let barrier = Arc::new(Barrier::new(num_threads));
        let results = Arc::new(Mutex::new(Vec::new()));

        let threads: Vec<_> = (0..num_threads)
            .map(|_| {
                let mut iter = iter.clone();
                let counter = Arc::clone(&counter);
                let barrier = Arc::clone(&barrier);
                let results = Arc::clone(&results);

                thread::spawn(move || {
                    barrier.wait();
                    let mut local_results = Vec::new();

                    for _ in 0..items_per_thread {
                        if let Some(item) = iter.next() {
                            counter.fetch_add(1, Ordering::SeqCst);
                            local_results.push(item);
                        }
                    }

                    results.lock().push(local_results);
                })
            })
            .collect();

        for thread in threads {
            thread.join().unwrap();
        }

        assert_eq!(counter.load(Ordering::SeqCst), total_items);

        let all_results = results.lock();
        let mut collected: Vec<_> = all_results.iter().flatten().copied().collect();
        collected.sort();
        assert_eq!(collected, (0..total_items).collect::<Vec<_>>());
    }

    #[test]
    fn test_block_on_stream_concurrent_clone_and_drive() {
        let num_items = 50;
        let num_threads = 3;

        let iter = CurrentThreadRuntime::new().block_on_stream_thread_safe(|h| {
            stream::unfold(0, move |state| {
                let h = h.clone();
                async move {
                    if state < num_items {
                        h.spawn_cpu(move || {
                            thread::sleep(Duration::from_micros(10));
                            state
                        })
                        .await;
                        Some((state, state + 1))
                    } else {
                        None
                    }
                }
            })
        });

        let collected = Arc::new(Mutex::new(Vec::new()));
        let barrier = Arc::new(Barrier::new(num_threads));

        let threads: Vec<_> = (0..num_threads)
            .map(|thread_id| {
                let iter = iter.clone();
                let collected = Arc::clone(&collected);
                let barrier = Arc::clone(&barrier);

                thread::spawn(move || {
                    barrier.wait();
                    let mut local_items = Vec::new();

                    for item in iter {
                        local_items.push((thread_id, item));
                        if local_items.len() >= 5 {
                            break;
                        }
                    }

                    collected.lock().extend(local_items);
                })
            })
            .collect();

        for thread in threads {
            thread.join().unwrap();
        }

        let results = collected.lock();
        let mut values: Vec<_> = results.iter().map(|(_, v)| *v).collect();
        values.sort();
        values.dedup();

        assert!(values.len() >= 5);
        assert!(values.iter().all(|&v| v < num_items));
    }

    #[test]
    fn test_block_on_stream_async_work() {
        let runtime = CurrentThreadRuntime::new();
        let handle = runtime.handle();
        let iter = runtime.block_on_stream({
            stream::unfold((handle, 0), |(h, state)| async move {
                if state < 10 {
                    let value = h
                        .spawn(async move { futures::future::ready(state * 2).await })
                        .await;
                    Some((value, (h, state + 1)))
                } else {
                    None
                }
            })
        });

        let results: Vec<_> = iter.collect();
        assert_eq!(results, vec![0, 2, 4, 6, 8, 10, 12, 14, 16, 18]);
    }

    #[test]
    fn test_block_on_stream_drop_receivers_early() {
        let counter = Arc::new(AtomicUsize::new(0));
        let c = Arc::clone(&counter);

        let mut iter = CurrentThreadRuntime::new().block_on_stream({
            stream::unfold(0, move |state| {
                let c = Arc::clone(&c);
                async move {
                    (state < 100).then(|| {
                        c.fetch_add(1, Ordering::SeqCst);
                        (state, state + 1)
                    })
                }
            })
            .boxed()
        });

        assert_eq!(iter.next(), Some(0));
        assert_eq!(iter.next(), Some(1));
        assert_eq!(iter.next(), Some(2));

        drop(iter);

        let final_count = counter.load(Ordering::SeqCst);
        assert!(
            final_count < 100,
            "Stream should stop when all receivers are dropped"
        );
    }

    #[test]
    fn test_block_on_stream_interleaved_access() {
        let barrier = Arc::new(Barrier::new(2));
        let iter = CurrentThreadRuntime::new()
            .block_on_stream_thread_safe(|_h| stream::iter(0..20).boxed());

        let iter1 = iter.clone();
        let iter2 = iter;
        let barrier1 = Arc::clone(&barrier);
        let barrier2 = barrier;

        let thread1 = thread::spawn(move || {
            let mut iter = iter1;
            let mut results = Vec::new();
            barrier1.wait();

            for _ in 0..5 {
                if let Some(val) = iter.next() {
                    results.push(val);
                    thread::sleep(Duration::from_micros(50));
                }
            }
            results
        });

        let thread2 = thread::spawn(move || {
            let mut iter = iter2;
            let mut results = Vec::new();
            barrier2.wait();

            for _ in 0..5 {
                if let Some(val) = iter.next() {
                    results.push(val);
                    thread::sleep(Duration::from_micros(50));
                }
            }
            results
        });

        let results1 = thread1.join().unwrap();
        let results2 = thread2.join().unwrap();

        let mut all_results = results1;
        all_results.extend(results2);
        all_results.sort();

        assert_eq!(all_results, (0..10).collect::<Vec<_>>());

        for i in 0..10 {
            assert_eq!(all_results.iter().filter(|&&x| x == i).count(), 1);
        }
    }

    #[test]
    fn test_block_on_stream_stress_test() {
        let num_threads = 10;
        let num_items = 1000;

        let iter = CurrentThreadRuntime::new()
            .block_on_stream_thread_safe(|_h| stream::iter(0..num_items).boxed());

        let received = Arc::new(Mutex::new(Vec::new()));
        let barrier = Arc::new(Barrier::new(num_threads));

        let threads: Vec<_> = (0..num_threads)
            .map(|_| {
                let iter = iter.clone();
                let received = Arc::clone(&received);
                let barrier = Arc::clone(&barrier);

                thread::spawn(move || {
                    barrier.wait();
                    for val in iter {
                        received.lock().push(val);
                    }
                })
            })
            .collect();

        for thread in threads {
            thread.join().unwrap();
        }

        let mut results = received.lock().clone();
        results.sort();

        assert_eq!(results.len(), num_items);
        assert_eq!(results, (0..num_items).collect::<Vec<_>>());
    }
}