Kotlin的协程：上下文

Kotlin 的协程上下文叫做 CoroutineContext，通常用来切换线程池。

CoroutineContext 的应用

launch

launch 的第一个参数 context 是 CoroutineContext，默认值是 EmptyCoroutineContext。

如果需要指定 launch 工作的线程池，就需要指定 CoroutineContext 参数。

public fun CoroutineScope.launch(
    context: CoroutineContext = EmptyCoroutineContext,
    start: CoroutineStart = CoroutineStart.DEFAULT,
    block: suspend CoroutineScope.() -> Unit
): Job {
    val newContext = newCoroutineContext(context)
    val coroutine = if (start.isLazy)
        LazyStandaloneCoroutine(newContext, block) else
        StandaloneCoroutine(newContext, active = true)
    coroutine.start(start, coroutine, block)
    return coroutine
}

withContext

withContext 用来切换线程执行代码。它的第一个参数是 CoroutineContext，指定线程池。

public suspend fun <T> withContext(
    context: CoroutineContext,
    block: suspend CoroutineScope.() -> T
): T {
}

在 getUserInfoIo 中指定 withContext 的上下文是 Dispatchers.IO。

fun main() = runBlocking {
    val user = getUserInfoIo()
    logX(user)
}

suspend fun getUserInfoIo(): String {
    logX("Before IO Context.")
    withContext(Dispatchers.IO) {
        logX("In IO Context.")
        delay(1000)
    }
    logX("After IO Context.")
    return "BoyCoder"
}

输出如下：

================================
Before IO Context.
Thread:main @coroutine#1, time:1656560405319
================================
================================
In IO Context.
Thread:DefaultDispatcher-worker-1 @coroutine#1, time:1656560405351
================================
================================
After IO Context.
Thread:main @coroutine#1, time:1656560406360
================================
================================
BoyCoder
Thread:main @coroutine#1, time:1656560406361
================================

可以看到在 Thread:DefaultDispatcher-worker-1 线程执行协程1。其他代码在 main 线程。

suspend main

main 函数有 suspend 关键字的版本，可以直接执行挂起函数。

suspend fun main() {
    val user = getUserInfoIo()
    logX(user)
}

它和 runBlocking 的区别在于 withContext 切换线程后，都执行在 DefaultDispatcher-worker-1 线程。

================================
Before IO Context.
Thread:main, time:1656569681374
================================
================================
In IO Context.
Thread:DefaultDispatcher-worker-1, time:1656569681440
================================
================================
After IO Context.
Thread:DefaultDispatcher-worker-1, time:1656569682449
================================
================================
BoyCoder
Thread:DefaultDispatcher-worker-1, time:1656569682449
================================

runBlocking

runBlocking 的第一个参数也是 CoroutineContext，默认为 EmptyCoroutineContext。

public fun <T> runBlocking(context: CoroutineContext = EmptyCoroutineContext, block: suspend CoroutineScope.() -> T): T {
}

可以给 runBlocking 增加自定义的 CoroutineContext。

fun main() = runBlocking(Dispatchers.IO) {
    val user = getUserInfoIo()
    logX(user)
}

输出如下：

================================
Before IO Context.
Thread:DefaultDispatcher-worker-2 @coroutine#1, time:1656570311862
================================
================================
In IO Context.
Thread:DefaultDispatcher-worker-2 @coroutine#1, time:1656570311896
================================
================================
After IO Context.
Thread:DefaultDispatcher-worker-2 @coroutine#1, time:1656570312903
================================
================================
BoyCoder
Thread:DefaultDispatcher-worker-2 @coroutine#1, time:1656570312903
================================

可以看出增加 Dispatchers.IO 后，协程一直执行在 DefaultDispatcher-worker-2 线程。

Dispatchers 调度器

Kotlin 内置的 Dispatchers 有 4 种，他们本质上都继承 CoroutineContext。

• Default：默认调度器，用于 CPU 密集型任务，线程数和核心数一致。
• Main：主线程调度器，只在 Android、Swing 等 UI 平台有用，普通 JVM 工程无法使用。
• UnConfined：无限制调度器，协程可能运行在任意线程上。
• IO：IO 调度器，用于 IO 密集型任务，线程数会多一些，比如 64 个线程

public actual object Dispatchers {
    /**
     * The default [CoroutineDispatcher] that is used by all standard builders like
     * [launch][CoroutineScope.launch], [async][CoroutineScope.async], etc
     * if no dispatcher nor any other [ContinuationInterceptor] is specified in their context.
     *
     * It is backed by a shared pool of threads on JVM. By default, the maximal level of parallelism used
     * by this dispatcher is equal to the number of CPU cores, but is at least two.
     * Level of parallelism X guarantees that no more than X tasks can be executed in this dispatcher in parallel.
     */
    @JvmStatic
    public actual val Default: CoroutineDispatcher = createDefaultDispatcher()

    /**
     * A coroutine dispatcher that is confined to the Main thread operating with UI objects.
     * This dispatcher can be used either directly or via [MainScope] factory.
     * Usually such dispatcher is single-threaded.
     *
     * Access to this property may throw [IllegalStateException] if no main thread dispatchers are present in the classpath.
     *
     * Depending on platform and classpath it can be mapped to different dispatchers:
     * - On JS and Native it is equivalent of [Default] dispatcher.
     * - On JVM it is either Android main thread dispatcher, JavaFx or Swing EDT dispatcher. It is chosen by
     *   [`ServiceLoader`](https://docs.oracle.com/javase/8/docs/api/java/util/ServiceLoader.html).
     *
     * In order to work with `Main` dispatcher, the following artifacts should be added to project runtime dependencies:
     *  - `kotlinx-coroutines-android` for Android Main thread dispatcher
     *  - `kotlinx-coroutines-javafx` for JavaFx Application thread dispatcher
     *  - `kotlinx-coroutines-swing` for Swing EDT dispatcher
     *
     * In order to set a custom `Main` dispatcher for testing purposes, add the `kotlinx-coroutines-test` artifact to 
     * project test dependencies.
     *
     * Implementation note: [MainCoroutineDispatcher.immediate] is not supported on Native and JS platforms.
     */
    @JvmStatic
    public actual val Main: MainCoroutineDispatcher get() = MainDispatcherLoader.dispatcher

    /**
     * A coroutine dispatcher that is not confined to any specific thread.
     * It executes initial continuation of the coroutine in the current call-frame
     * and lets the coroutine resume in whatever thread that is used by the corresponding suspending function, without
     * mandating any specific threading policy. Nested coroutines launched in this dispatcher form an event-loop to avoid
     * stack overflows.
     *
     * ### Event loop
     * Event loop semantics is a purely internal concept and have no guarantees on the order of execution
     * except that all queued coroutines will be executed on the current thread in the lexical scope of the outermost
     * unconfined coroutine.
     *
     * For example, the following code:
     * ```
     * withContext(Dispatchers.Unconfined) {
     *    println(1)
     *    withContext(Dispatchers.Unconfined) { // Nested unconfined
     *        println(2)
     *    }
     *    println(3)
     * }
     * println("Done")
     * ```
     * Can print both "1 2 3" and "1 3 2", this is an implementation detail that can be changed.
     * But it is guaranteed that "Done" will be printed only when both `withContext` are completed.
     *
     *
     * Note that if you need your coroutine to be confined to a particular thread or a thread-pool after resumption,
     * but still want to execute it in the current call-frame until its first suspension, then you can use
     * an optional [CoroutineStart] parameter in coroutine builders like
     * [launch][CoroutineScope.launch] and [async][CoroutineScope.async] setting it to the
     * the value of [CoroutineStart.UNDISPATCHED].
     */
    @JvmStatic
    public actual val Unconfined: CoroutineDispatcher = kotlinx.coroutines.Unconfined

    /**
     * The [CoroutineDispatcher] that is designed for offloading blocking IO tasks to a shared pool of threads.
     *
     * Additional threads in this pool are created and are shutdown on demand.
     * The number of threads used by this dispatcher is limited by the value of
     * "`kotlinx.coroutines.io.parallelism`" ([IO_PARALLELISM_PROPERTY_NAME]) system property.
     * It defaults to the limit of 64 threads or the number of cores (whichever is larger).
     *
     * Moreover, the maximum configurable number of threads is capped by the
     * `kotlinx.coroutines.scheduler.max.pool.size` system property.
     * If you need a higher number of parallel threads,
     * you should use a custom dispatcher backed by your own thread pool.
     *
     * This dispatcher shares threads with a [Default][Dispatchers.Default] dispatcher, so using
     * `withContext(Dispatchers.IO) { ... }` does not lead to an actual switching to another thread —
     * typically execution continues in the same thread.
     */
    @JvmStatic
    public val IO: CoroutineDispatcher = DefaultScheduler.IO
}

Dispatchers.IO

Dispatchers.IO 可能会复用 Dispatchers.Default 的线程。从上面例子可以看出，虽然设置的 Dispatchers.IO，实际是 DefaultDispatcher-worker 线程。

将 runBlocking 的 CoroutineContext 改为 Dispatchers.Default

fun main() = runBlocking(Dispatchers.Default) {
    val user = getUserInfoIo()
    logX(user)
}

输出结果如下：

================================
Before IO Context.
Thread:DefaultDispatcher-worker-1 @coroutine#1, time:1656575072324
================================
================================
In IO Context.
Thread:DefaultDispatcher-worker-1 @coroutine#1, time:1656575072358
================================
================================
After IO Context.
Thread:DefaultDispatcher-worker-1 @coroutine#1, time:1656575073366
================================
================================
BoyCoder
Thread:DefaultDispatcher-worker-1 @coroutine#1, time:1656575073367
================================

可以看到 withContext 切换到 IO 后，也用了 DefaultDispatcher-worker-1。这是因为 Dispatchers.Default 被 Dispatchers.IO 复用线程。

自定义 Dispatchers

使用自定义的 executor 然后转换为 Dispatcher 作为 CoroutineContext。

val mySingleDispatcher = Executors.newSingleThreadExecutor {
    Thread(it, "mySingleThread").apply {
        isDaemon = true
    }
}
    .asCoroutineDispatcher()

fun main() = runBlocking(mySingleDispatcher) {
    val user = getUserInfoIo()
    logX(user)
}

输出如下：

================================
Before IO Context.
Thread:mySingleThread @coroutine#1, time:1656575809476
================================
================================
In IO Context.
Thread:DefaultDispatcher-worker-1 @coroutine#1, time:1656575809510
================================
================================
After IO Context.
Thread:mySingleThread @coroutine#1, time:1656575810521
================================
================================
BoyCoder
Thread:mySingleThread @coroutine#1, time:1656575810521
================================

只有 In IO Context 运行在 DefaultDispatcher-worker-1，其他代码都运行在自定义的 dispatcher。

Dispatchers.Unconfined

如果 launch 使用默认 context，执行顺序为 1、4、2、3。

fun main() = runBlocking {
    logX("Before launch") // 1
    launch {
        logX("In launch") // 2
        delay(1000)
        logX("End launch") // 3
    }
    logX("After launch") // 4
}

输出如下

================================
Before launch
Thread:main @coroutine#1, time:1656576229645
================================
================================
After launch
Thread:main @coroutine#1, time:1656576229677
================================
================================
In launch
Thread:main @coroutine#2, time:1656576229679
================================
================================
End launch
Thread:main @coroutine#2, time:1656576230686
================================

如果 launch 使用 Unconfined，执行顺序是不定的。

fun main() = runBlocking {
    logX("Before launch") // 1
    launch(Dispatchers.Unconfined) {
        logX("In launch") // 2
        delay(1000)
        logX("End launch") // 3
    }
    logX("After launch") // 4
}

输出如下

================================
Before launch
Thread:main @coroutine#1, time:1656576759031
================================
================================
In launch
Thread:main @coroutine#2, time:1656576759055
================================
================================
After launch
Thread:main @coroutine#1, time:1656576759060
================================
================================
End launch
Thread:kotlinx.coroutines.DefaultExecutor @coroutine#2, time:1656576760059
================================

执行顺序变为 1、2、4、3。而且标记 3 执行在 DefaultExecutor。

Dispatchers.Unconfined 可能执行在任何线程，不应随意使用 Dispatchers.Unconfined。

CoroutineScope 协程作用域

使用 launch 时，必须有 CoroutineScope 协程作用域，launch 是 CoroutineScope 的扩展函数。

CoroutineScope 的实现很简单，它是接口类，只有 coroutineContext 成员。

public interface CoroutineScope {
    /**
     * The context of this scope.
     * Context is encapsulated by the scope and used for implementation of coroutine builders that are extensions on the scope.
     * Accessing this property in general code is not recommended for any purposes except accessing the [Job] instance for advanced usages.
     *
     * By convention, should contain an instance of a [job][Job] to enforce structured concurrency.
     */
    public val coroutineContext: CoroutineContext
}

指定自定义的 CoroutineScope，使用 scope 启动 3 个协程，实现结构化并发。

fun main() = runBlocking {
    val scope = CoroutineScope(Job())

    scope.launch {
        logX("first start")
        delay(1000)
        logX("first end")
    }

    scope.launch {
        logX("second start")
        delay(1000)
        logX("second end")
    }

    scope.launch {
        logX("third start")
        delay(1000)
        logX("third end")
    }

    delay(500)
    scope.cancel()
    delay(1000)
}

输出如下

================================
first start
Thread:DefaultDispatcher-worker-2 @coroutine#2, time:1656577491658
================================
================================
third start
Thread:DefaultDispatcher-worker-3 @coroutine#4, time:1656577491663
================================
================================
second start
Thread:DefaultDispatcher-worker-1 @coroutine#3, time:1656577491660
================================

所有 scope 启动的协程都被取消，没有继续执行。

Job 和 Dispatcher

Job

Job 继承了 CoroutineContext.Element。

public interface Job : CoroutineContext.Element {
}

Element 继承 CoroutineContext

@kotlin.SinceKotlin public interface CoroutineContext {
    public interface Element : kotlin.coroutines.CoroutineContext {
    }
}

因此 Job 本身就是 CoroutineContext。

Dispatchers

Dispatchers 是单例类，内部的有 4 个预置 Dispatchers。

public actual object Dispatchers {
}

以 Default 为例，它是 CoroutineDispatcher 类，继承关系如下：
CoroutineDispatcher -> ContinuationInterceptor -> CoroutineContext.Element -> CoroutineContext

因此 Dispatchers 也是 CoroutineContext。

CoroutineContext 的设计

CoroutineContext 的设计方式和 Map 的设计方式很类似。

public interface CoroutineContext {
    public operator fun <E : Element> get(key: Key<E>): E?

    public fun <R> fold(initial: R, operation: (R, Element) -> R): R

    public operator fun plus(context: CoroutineContext): CoroutineContext =

    public fun minusKey(key: Key<*>): CoroutineContext

• get 相当于 map 的 get
• fold 相当于 map 的 foreach
• plus 相当于 map 的 put
• minusKey 相当于 map 的 remove

利用 CoroutineContext 的接口，可以写出 + 、[] 这种类似集合的操作。因为 CoroutineContext 重载了 plus 操作符，可以用 + 代替 plus；重载 get 操作符，可以用 [] 代替 get。

@OptIn(ExperimentalStdlibApi::class)
fun main() = runBlocking {
    val scope = CoroutineScope(Job() + mySingleDispatcher)

    scope.launch {
        logX(coroutineContext[CoroutineDispatcher] == mySingleDispatcher)
        delay(500)
        logX("first end")
    }

    delay(500)
    scope.cancel()
    delay(1000)
}

输出如下

================================
true
Thread:mySingleThread @coroutine#2, time:1656584502696
================================

可以看出 CoroutineContext 内部的 dispatchers 就是 mySingleDispatcher。

CoroutineName 协程名称

CoroutineName 可以指定协程名称，它本质也是 CoroutineContext。

@OptIn(ExperimentalStdlibApi::class)
fun main() = runBlocking {
    val scope = CoroutineScope(Job() + mySingleDispatcher)

    scope.launch(CoroutineName("MyFirstCoroutine")) {
        logX(coroutineContext[CoroutineDispatcher] == mySingleDispatcher)
        delay(500)
        logX("first end")
    }

    delay(500)
    scope.cancel()
    delay(1000)
}

输出如下

================================
true
Thread:mySingleThread @MyFirstCoroutine#2, time:1656584886943
================================

协程名称变为 CoroutineName 定义的 MyFirstCoroutine，#2 是协程 id。

CoroutineExceptionHandler 协程异常处理

CoroutineExceptionHandler 用来捕获协程中的异常，它也是 CoroutineContext。

@OptIn(ExperimentalStdlibApi::class)
suspend fun main() {
    val scope = CoroutineScope(Job() + mySingleDispatcher)
    val handler = CoroutineExceptionHandler { _, throwable ->
        println("catch exception $throwable")
    }
    val job = scope.launch(handler) {
        logX(coroutineContext[CoroutineDispatcher] == mySingleDispatcher)
        val str: String? = null
        str!!.length
        logX("first end")
    }
    job.join()
}

输出如下

================================
true
Thread:mySingleThread @coroutine#1, time:1656585762125
================================
catch exception java.lang.NullPointerException

故意抛出空指针异常，然后给 CoroutineExceptionHandler 捕获。

挂起函数和 CoroutineContext

suspend 挂起函数能直接访问 coroutineContext。

suspend fun testCoroutineContext() = coroutineContext

fun main() = runBlocking {
    print(testCoroutineContext())
}

输出如下

[CoroutineId(1), "coroutine#1":BlockingCoroutine{Active}@63e2203c, BlockingEventLoop@1efed156]

挂起函数需要在协程或者另外一个挂起函数中运行，因此它也能访问协程的 CoroutineContext。

coroutineContext 是 Continuation.kt 的内部成员。

public suspend inline val coroutineContext: CoroutineContext
    get() {
        throw NotImplementedError("Implemented as intrinsic")
    }

总结

• CoroutineContext 协程上下文和 map 很类似，和 map 一样，它可以添加或者获取 Element。
• Job、Dispatchers、CoroutineName、CoroutineExceptionHandler 本质都是 CoroutineContext。
• CoroutineScope 封装了 CoroutineContext，CoroutineContext 是 CoroutineScope 的成员。
• suspend 挂起函数也和 CoroutineContext 有关。挂起函数要在协程中运行，协程有它的 CoroutineContext，因此挂起函数也能访问 CoroutineContext。