我愿称之为最好用的golang并发控制库

Concurrency

install

GOPROXY=https://goproxy.cn go get -v github.com/lxzan/concurrency@latest

Feature

• 并发协程数量限制
• 支持 contex.Contex
• 支持 panic recover, 返回包含错误堆栈的 error
• 递归实现的任务调度, 不依赖 time.Ticker 和 channel

Usage

• WorkerGroup 工作组, 添加一组任务, 等待执行完成, 可以很好的替代WaitGroup.

package main

import (
    "fmt"
    "github.com/lxzan/concurrency"
    "sync/atomic"
)

func main() {
    sum := int64(0)
    w := concurrency.NewWorkerGroup()
    for i := int64(1); i <= 10; i++ {
        w.AddJob(concurrency.Job{
            Args: i,
            Do: func(args interface{}) error {
                fmt.Printf("%v ", args)
                atomic.AddInt64(&sum, args.(int64))
                return nil
            },
        })
    }
    w.StartAndWait()
    fmt.Printf("sum=%d\n", sum)
}

4 5 6 7 8 9 10 1 3 2 sum=55

• WorkerQueue 工作队列, 可以不断往里面添加任务, 一旦有CPU资源空闲就去执行

package main

import (
    "fmt"
    "github.com/lxzan/concurrency"
    "time"
)

func Add(args interface{}) error {
    arr := args.([]int)
    ans := 0
    for _, item := range arr {
        ans += item
    }
    fmt.Printf("args=%v, ans=%d\n", args, ans)
    return nil
}

func Mul(args interface{}) error {
    arr := args.([]int)
    ans := 1
    for _, item := range arr {
        ans *= item
    }
    fmt.Printf("args=%v, ans=%d\n", args, ans)
    return nil
}

func main() {
    args1 := []int{1, 3}
    args2 := []int{1, 3, 5}
    w := concurrency.NewWorkerQueue()
    w.AddJob(
        concurrency.Job{Args: args1, Do: Add},
        concurrency.Job{Args: args1, Do: Mul},
        concurrency.Job{Args: args2, Do: Add},
        concurrency.Job{Args: args2, Do: Mul},
    )
    w.StopAndWait(30*time.Second)
}

args=[1 3], ans=4
args=[1 3 5], ans=15
args=[1 3], ans=3
args=[1 3 5], ans=9