前言
在 Go 1.7 版本之前,context 还是非编制的,它存在于 golang.org/x/net/context 包中。
后来,Golang 团队发现 context 还挺好用的,就把 context 收编了,在 Go 1.7 版本正式纳入了标准库。
为什么需要Context
当一个协程(goroutine)开启后,我们是无法强制关闭它的。
常见的关闭协程的原因有如下几种:
goroutine 自己跑完结束退出(正常关闭,本文不讨论)。
主进程 crash 退出,goroutine 被迫退出(属异常关闭,应优化代码)。
通过通道发送信号,引导协程的关闭(属于开发者手动控制协程的方法)。
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 func main () { stopSingal := make (chan bool ) for i := 1 ; i <= 5 ; i++ { go monitor(stopSingal, i) } time.Sleep(1 * time.Second) close (stopSingal) time.Sleep(5 * time.Second) fmt.Println("主程序退出!!" ) } func monitor (ch chan bool , number int ) { for { select { case v := <-ch: fmt.Printf("监控器%v,接收到通道值为:%v,监控结束。\n" , number, v) return default : fmt.Printf("监控器%v,正在监控中...\n" , number) time.Sleep(2 * time.Second) } } }
1 2 3 4 5 6 7 8 9 10 11 监控器1,正在监控中... 监控器5,正在监控中... 监控器4,正在监控中... 监控器2,正在监控中... 监控器3,正在监控中... 监控器2,接收到通道值为:false,监控结束。 监控器1,接收到通道值为:false,监控结束。 监控器3,接收到通道值为:false,监控结束。 监控器5,接收到通道值为:false,监控结束。 监控器4,接收到通道值为:false,监控结束。 主程序退出!!
上面的例子,在我们定义一个无缓冲通道时,要实现对所有的 goroutine 进行关闭,可以使用 close 关闭通道,然后在所有的 goroutine 里不断检查通道是否关闭*(前提你得约定好,该通道你只会进行 close 而不会发送其他数据,否则发送一次数据就会关闭一个goroutine,这样会不符合咱们的预期,所以最好你对这个通道再做一层封装做个限制)*来决定是否结束 goroutine。
所以你看到这里,我做为初学者还是没有找到使用 Context 的必然理由,我只能说 Context 是个很好用的东西,使用它方便了我们在处理并发时候的一些问题,但是它并不是不可或缺的。
换句话说,它解决的并不是 能不能 的问题,而是解决 更好用 的问题。
简单使用Context
此处的代码,我们先实现一个简单的 Context Demo(我使用 Context 对上面的例子进行了一番改造),然后详细分析其中的代码意义。
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 func main () { ctx, cancel := context.WithCancel(context.Background()) for i :=1 ; i <= 5 ; i++ { go monitor(ctx, i) } time.Sleep( 1 * time.Second) cancel() time.Sleep( 5 * time.Second) fmt.Println("主程序退出!!" ) } func monitor (ctx context.Context, number int ) { for { select { case v :=<- ctx.Done(): fmt.Printf("监控器%v,接收到通道值为:%v,监控结束。\n" , number,v) return default : fmt.Printf("监控器%v,正在监控中...\n" , number) time.Sleep(2 * time.Second) } } }
代码分析
1 ctx, cancel := context.WithCancel(context.Background())
以 context.Background() 为 parent context 定义一个可取消的 context
然后你可以在所有的goroutine 里利用 for + select 搭配来不断检查 ctx.Done() 是否可读,可读就说明该 context 已经取消,你可以清理 goroutine 并退出了。
当你想到取消 context 的时候,只要调用一下 cancel 方法即可。这个 cancel 就是我们在创建 ctx 的时候返回的第二个值。
该程序的运行结果如下 :
1 2 3 4 5 6 7 8 9 10 监控器3 ,正在监控中... 监控器4 ,正在监控中... 监控器1 ,正在监控中... 监控器2 ,正在监控中... 监控器2 ,接收到通道值为:{},监控结束。 监控器5 ,接收到通道值为:{},监控结束。 监控器4 ,接收到通道值为:{},监控结束。 监控器1 ,接收到通道值为:{},监控结束。 监控器3 ,接收到通道值为:{},监控结束。 主程序退出!!
根Context 是什么?
创建 Context 必须要指定一个 父 Context,当我们要创建第一个Context时该怎么办呢?
不用担心,Go 已经帮我们实现了2个,我们代码中最开始都是以这两个内置context作为最顶层的parent context,衍生出更多的子Context。
1 2 3 4 5 6 7 8 9 10 11 12 var ( background = new (emptyCtx) todo = new (emptyCtx) ) func Background () Context { return background } func TODO () Context { return todo }
Background :主要用于main函数、初始化以及测试代码中,作为Context这个树结构的最顶层的Context,也就是根Context,它不能被取消。
TODO :如果我们不知道该使用什么Context的时候,可以使用这个,但是实际应用中,暂时还没有使用过这个TODO。
他们两个本质上都是emptyCtx结构体类型,是一个不可取消,没有设置截止时间,没有携带任何值的Context。
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 type emptyCtx int func (*emptyCtx) Deadline() (deadline time.Time, ok bool ) { return } func (*emptyCtx) Done() <-chan struct {} { return nil } func (*emptyCtx) Err() error { return nil } func (*emptyCtx) Value(key interface {}) interface {} { return nil }
Context 的继承衍生
上面在定义我们自己的 Context 时,我们使用的是 WithCancel 这个方法。除它之外,context 包还有其他几个 With 系列的函数.
1 2 3 4 func WithCancel (parent Context) (ctx Context, cancel CancelFunc)func WithDeadline (parent Context, deadline time.Time) (Context, CancelFunc)func WithTimeout (parent Context, timeout time.Duration) (Context, CancelFunc)func WithValue (parent Context, key, val interface {}) Context
这四个函数有一个共同的特点,就是第一个参数,都是接收一个 父context。
WithCancel
WithCancel 返回带有新Done通道的父节点的副本。当调用返回的cancel函数或当关闭父上下文的Done通道时,将关闭返回上下文的Done通道,无论先发生什么情况。
取消此上下文将释放与其关联的资源,因此代码应该在此上下文中运行的操作完成后立即调用cancel。
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 func main () { ctx, cancel := context.WithCancel(context.Background()) for i := 1 ; i <= 5 ; i++ { go job1(ctx, i) } time.Sleep(1 * time.Second) cancel() time.Sleep(5 * time.Second) fmt.Println("主程序退出!!" ) } func job1 (ctx context.Context, number int ) { for { select { case v := <-ctx.Done(): fmt.Printf("监控器%v,接收到通道值为:%v,监控结束。\n" , number, v) return default : fmt.Printf("监控器%v,正在监控中...\n" , number) time.Sleep(2 * time.Second) } } }
1 2 3 4 5 6 7 8 9 10 11 监控器5 ,正在监控中... 监控器2 ,正在监控中... 监控器3 ,正在监控中... 监控器4 ,正在监控中... 监控器1 ,正在监控中... 监控器2 ,接收到通道值为:{},监控结束。 监控器4 ,接收到通道值为:{},监控结束。 监控器5 ,接收到通道值为:{},监控结束。 监控器1 ,接收到通道值为:{},监控结束。 监控器3 ,接收到通道值为:{},监控结束。 主程序退出!!
WithDeadline
WithDeadline 返回父上下文的副本,并将deadline调整为不迟于d。如果父上下文的deadline已经早于d,则WithDeadline(parent, d)在语义上等同于父上下文。当截止日过期时,当调用返回的cancel函数时,或者当父上下文的Done通道关闭时,返回上下文的Done通道将被关闭,以最先发生的情况为准。
取消此上下文将释放与其关联的资源,因此代码应该在此上下文中运行的操作完成后立即调用cancel。
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 func main () { ctx01, cancel := context.WithCancel(context.Background()) ctx02, cancel := context.WithDeadline(ctx01, time.Now().Add(1 * time.Second)) defer cancel() for i :=1 ; i <= 5 ; i++ { go job2(ctx02, i) } time.Sleep(5 * time.Second) if ctx02.Err() != nil { fmt.Println("监控器取消的原因: " , ctx02.Err()) } fmt.Println("主程序退出!!" ) } func job2 (ctx context.Context, number int ) { for { select { case <- ctx.Done(): fmt.Printf("监控器%v,监控结束。\n" , number) return default : fmt.Printf("监控器%v,正在监控中...\n" , number) time.Sleep(2 * time.Second) } } }
1 2 3 4 5 6 7 8 9 10 11 12 监控器5 ,正在监控中... 监控器3 ,正在监控中... 监控器4 ,正在监控中... 监控器2 ,正在监控中... 监控器1 ,正在监控中... 监控器1 ,监控结束。 监控器3 ,监控结束。 监控器2 ,监控结束。 监控器4 ,监控结束。 监控器5 ,监控结束。 监控器取消的原因: context deadline exceeded 主程序退出!!
WithTimeout
WithTimeout 返回WithDeadline(parent, time.Now().Add(timeout))。
取消此上下文将释放与其相关的资源,因此代码应该在此上下文中运行的操作完成后立即调用cancel,通常用于数据库或者网络连接的超时控制。
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 func main () { ctx01, cancel := context.WithCancel(context.Background()) ctx02, cancel := context.WithTimeout(ctx01, 1 * time.Second) defer cancel() for i :=1 ; i <= 5 ; i++ { go job3(ctx02, i) } time.Sleep(5 * time.Second) if ctx02.Err() != nil { fmt.Println("监控器取消的原因: " , ctx02.Err()) } fmt.Println("主程序退出!!" ) } func job3 (ctx context.Context, number int ) { for { select { case <- ctx.Done(): fmt.Printf("监控器%v,监控结束。\n" , number) return default : fmt.Printf("监控器%v,正在监控中...\n" , number) time.Sleep(2 * time.Second) } } }
1 2 3 4 5 6 7 8 9 10 11 12 监控器5 ,正在监控中... 监控器4 ,正在监控中... 监控器2 ,正在监控中... 监控器3 ,正在监控中... 监控器1 ,正在监控中... 监控器4 ,监控结束。 监控器1 ,监控结束。 监控器2 ,监控结束。 监控器5 ,监控结束。 监控器3 ,监控结束。 监控器取消的原因: context deadline exceeded 主程序退出!!
WithValue
WithValue 返回父节点的副本,其中与key关联的值为val。
仅对API和进程间传递请求域的数据使用上下文值,而不是使用它来传递可选参数给函数。
所提供的键必须是可比较的,并且不应该是string类型或任何其他内置类型,以避免使用上下文在包之间发生冲突。WithValue的用户应该为键定义自己的类型。为了避免在分配给interface{}时进行分配,上下文键通常具有具体类型struct{}。或者,导出的上下文关键变量的静态类型应该是指针或接口。
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 func main () { ctx01, cancel := context.WithCancel(context.Background()) ctx02, cancel := context.WithTimeout(ctx01, 1 * time.Second) ctx03 := context.WithValue(ctx02, "item" , "CPU" ) defer cancel() for i :=1 ; i <= 5 ; i++ { go job4(ctx03, i) } time.Sleep(5 * time.Second) if ctx02.Err() != nil { fmt.Println("监控器取消的原因: " , ctx02.Err()) } fmt.Println("主程序退出!!" ) } func job4 (ctx context.Context, number int ) { for { select { case <- ctx.Done(): fmt.Printf("监控器%v,监控结束。\n" , number) return default : value := ctx.Value("item" ) fmt.Printf("监控器%v,正在监控 %v \n" , number, value) time.Sleep(2 * time.Second) } } }
1 2 3 4 5 6 7 8 9 10 11 12 监控器5 ,正在监控 CPU 监控器1 ,正在监控 CPU 监控器4 ,正在监控 CPU 监控器3 ,正在监控 CPU 监控器2 ,正在监控 CPU 监控器3 ,监控结束。 监控器1 ,监控结束。 监控器5 ,监控结束。 监控器4 ,监控结束。 监控器2 ,监控结束。 监控器取消的原因: context deadline exceeded 主程序退出!!