go - 为什么在主goroutine终止后subgoroutine可以运行？

Question

为什么gogoroutine在主goroutine终止后可以运行？
这是代码

package main

import (
    "fmt"
    "log"
    "time"
)


func main() {
    ch := make(chan int)
    log.Println("===========")
    go func() {
        log.Println("doing in sub goroutine")
        ch <- 1
        log.Println("done in sub goroutine") 
    }()

    // mock slowly proceedings
    time.Sleep(1 * time.Second)

    log.Println("doing in main goroutine")
    fmt.Println("x in ch is:", <-ch)      
    log.Println("done in main goroutine") 



    // output：
    // 2020/07/04 16:26:35 doing in sub goroutine
    // 2020/07/04 16:26:36 doing in main goroutine
    // x in ch is: 1
    // 2020/07/04 16:26:36 done in main goroutine 
    // 2020/07/04 16:26:36 done in sub goroutine (should not be output)

}

当subgoroutine在主goroutine终端之后可以运行时，以下输出很奇怪。
//2020/07/04 16:26:36在主goroutine中完成
//在子goroutine中完成2020/07/04 16:26:36(不应输出)

Answer 1

打印日志消息后，主goroutine仍在运行。另一个goroutine仍然可以运行，直到主函数返回为止，此后再运行一点，直到运行库完成其清理为止。