用来管理goroutine的工具,执行一个task就调度一个worker来运行,然后对worker进行管理,一个worker也相当于一个goroutine.
ants用一个pool管理worker,负责worker的调度还有回收
type Pool struct { // capacity of the pool. capacity int32 // running is the number of the currently running goroutines. running int32 // expiryDuration set the expired time (second) of every worker. expiryDuration time.Duration // workers is a slice that store the available workers. workers []*goWorker // release is used to notice the pool to closed itself. release int32 // lock for synchronous operation. lock sync.Mutex // cond for waiting to get a idle worker. cond *sync.Cond // once makes sure releasing this pool will just be done for one time. once sync.Once // workerCache speeds up the obtainment of the an usable worker in function:retrieveWorker. workerCache sync.Pool // panicHandler is used to handle panics from each worker goroutine. // if nil, panics will be thrown out again from worker goroutines. panicHandler func(interface{}) // Max number of goroutine blocking on pool.Submit. // 0 (default value) means no such limit. maxBlockingTasks int32 // goroutine already been blocked on pool.Submit // protected by pool.lock blockingNum int32 // When nonblocking is true, Pool.Submit will never be blocked. // ErrPoolOverload will be returned when Pool.Submit cannot be done at once. // When nonblocking is true, MaxBlockingTasks is inoperative. nonblocking bool}type goWorker struct { // pool who owns this worker. pool *Pool // task is a job should be done. task chan func() // recycleTime will be update when putting a worker back into queue. recycleTime time.Time}
从pool中获取一个可用worker
pool -> retrieveWorker():
判断p.workers是否有闲置的worker可以取,有的话,从最后取出一个worker.
若p.workers没有闲置的worker可以取,则判断运行中的worker是否达到上限, 若没有,从p.workerCache取出一个woker,若p.workerCache没有就新建一个worker.
若运行中的worker达到上限,且是不阻塞的pool则直接返回nil.
若运行中的worker达到上限,且是阻塞的pool,则再阻塞数量没超过限定的数量时,p.blockingNum++并等待信号.
接收到信号的时候p.blockingNum—,判断p.workers中是否有可以运行的worker,有则返回worker,若没有则重新阻塞.
释放pool里的worker
pool -> Release():
只执行一次
p.release置1
所有worker都传nil task
调整pool中worker数量:
tune(size uint)
除了pool,worker.但确认执行的task中的func都是一样时,可用pool_func,worker_func.其中传递不是task了,而是task执行的参数.worker获取到参数,就塞进默认的func来执行.
