本实验使用的是Command Line Tool程序，并非IOS APPlication
异步调用
DispatchQueue.global().async {
print("子线程调用" + Thread.current.description + "!")
}
print("主线程调用" + Thread.current.description)
输出
主线程调用<NSThread: 0x1012033b0>{number = 1, name = main}
子线程调用<NSThread: 0x100f4dda0>{number = 2, name = (null)}!
异步调用获取数据回到主线程更新UI
DispatchQueue.global().async {
print("子线程调用" + Thread.current.description + "!") //这里同步调用了
DispatchQueue.main.async {
print("主线程更新UI" + Thread.current.description)
}
}
print("主线程调用" + Thread.current.description)
RunLoop.current.run()
由于程序使用Command Line Tool程序 主线程没有RunLoop 所以加上 RunLoop.current.run()
在IOS开发中不用添加
输出
主线程调用<NSThread: 0x100f035c0>{number = 1, name = main}
子线程调用<NSThread: 0x101838e00>{number = 2, name = (null)}!
主线程更新UI<NSThread: 0x100f035c0>{number = 1, name = main}
可见异步开启了新的线程
异步请求网络数据返回后更新本地
上面的例子是在异步中同步调用数据，一般我们网络数据请求接口封装的都是异步，我们以文件下载为例 (默认使用全局队列)
class LoadHelper {
//��模拟异步下载文件
static func loadFile(_ name:String,_ queue:DispatchQueue = DispatchQueue.global() , _ complete:@escaping ()->Void)
{
queue.async {
print("loading " + name + " in " + Thread.current.description)
let time = arc4random()%3 + 1
sleep(time)
complete()
}
}
}
异步下载一个文件
在main.swift中
LoadHelper.loadFile("1.jpg") {
print("1.jpg" + " load success....")
}
RunLoop.current.run()
输出
loading 1.jpg in <NSThread: 0x100f8b380>{number = 2, name = (null)}
1.jpg load success....
异步下载多个文件
let array = ["1.jpg", "2.jpg", "3.jpg"]
array.forEach { (name:String) in
LoadHelper.loadFile(name) {
print(name + " load success....")
}
}
RunLoop.current.run()
输出
loading 2.jpg in <NSThread: 0x102103cc0>{number = 4, name = (null)}
loading 1.jpg in <NSThread: 0x100f257f0>{number = 2, name = (null)}
loading 3.jpg in <NSThread: 0x10203c9c0>{number = 3, name = (null)}
2.jpg load success....
1.jpg load success....
3.jpg load success....
这个实验的输出前三行或后三行顺序有时会不一样
从输出中可以看除这个实验中，每次下载都开启了一个线程，三个下载任务并发执行的，各个之间没有联系，并且我们不知道什么时候全部下载完毕
多文件下载案例应用
比如现在有这么一个需求，并发下载多个文件，下载完后打印全部下载成功
方法一 （设置下载标记，每个任务下载完后检查标记）
let array = ["1.jpg", "2.jpg", "3.jpg"]
let lock = NSLock()
var completeNum = 0
array.forEach { (name:String) in
LoadHelper.loadFile(name) {
print(name + " load success....")
lock.lock()
completeNum = completeNum + 1
if completeNum == array.count
{
print("loading all success!!!")
print("Splice all Image!!!")
}
lock.unlock()
}
}
RunLoop.current.run()
输出
loading 3.jpg in <NSThread: 0x100f06040>{number = 2, name = (null)}
loading 1.jpg in <NSThread: 0x10191e610>{number = 4, name = (null)}
loading 2.jpg in <NSThread: 0x101a13c60>{number = 3, name = (null)}
1.jpg load success....
2.jpg load success....
3.jpg load success....
loading all success!!!
这种方法比较Low,一般不会这么做
方法二 (栅栏)
let array = ["1.jpg", "2.jpg", "3.jpg"]
let queue = DispatchQueue(label: "anson.gcd.com", attributes: .concurrent)
array.forEach { (name:String) in
LoadHelper.loadFile(name, queue) {
print(name + " load success....")
}
}
let wirte = DispatchWorkItem(flags: .barrier) {
print("loading all success!!!")
}
queue.async(execute: wirte)
print("---------------")
RunLoop.current.run()
输出
---------------
loading 3.jpg in <NSThread: 0x100f68010>{number = 4, name = (null)}
loading 2.jpg in <NSThread: 0x101a00b00>{number = 3, name = (null)}
loading 1.jpg in <NSThread: 0x100f67f00>{number = 2, name = (null)}
2.jpg load success....
3.jpg load success....
1.jpg load success....
loading all success!!!
这种方式比较容易理解 而且是非租塞的形式 一般可以使用这种方式获得多异步任务的整体结果回调
方法三 (group)
let array = ["1.jpg", "2.jpg", "3.jpg"]
let queue = DispatchQueue(label: "anson.gcd.com", attributes: .concurrent)
let group = DispatchGroup()
array.forEach { (name:String) in
group.enter()
LoadHelper.loadFile(name, queue) {
print(name + " load success....")
group.leave()
}
}
group.notify(queue: queue) {
print("loading all success!!!")
}
print("---------------")
RunLoop.current.run()
输出
---------------
loading 1.jpg in <NSThread: 0x102a64e90>{number = 4, name = (null)}
loading 3.jpg in <NSThread: 0x102a64610>{number = 2, name = (null)}
loading 2.jpg in <NSThread: 0x102907eb0>{number = 3, name = (null)}
2.jpg load success....
1.jpg load success....
3.jpg load success....
loading all success!!!
这种也是比较常用的
方法四 （GCD 信号量）
let array = ["1.jpg", "2.jpg", "3.jpg"]
let queue = DispatchQueue(label: "anson.gcd.com", attributes: .concurrent)
let semaphoreSignal = DispatchSemaphore(value:0)
array.forEach { (name:String) in
LoadHelper.loadFile(name, queue) {
print(name + " load success....")
semaphoreSignal.signal()
}
}
array.forEach { (name:String) in
semaphoreSignal.wait()
}
print("loading all success!!!")
print("---------------")
RunLoop.current.run()
输出
loading 2.jpg in <NSThread: 0x1021219f0>{number = 3, name = (null)}
loading 3.jpg in <NSThread: 0x1021219b0>{number = 2, name = (null)}
loading 1.jpg in <NSThread: 0x100f43850>{number = 4, name = (null)}
2.jpg load success....
3.jpg load success....
1.jpg load success....
loading all success!!!
---------------
虽然可以实现，但是最后是阻塞的，前面的异步任务没执行完，后面就走不过去
多任务并发控制最大线程数
由于并发任务每一个任务会创建对应的线程，所以有时候我们需要控制任务个数
我们可以在上面方法二到基础上加上信号量控制
let array = ["1.jpg", "2.jpg", "3.jpg", "4.jpg", "5.jpg", "6.jpg"]
let queue = DispatchQueue(label: "anson.gcd.com", attributes: .concurrent)
let semaphoreSignal = DispatchSemaphore(value:2)
array.forEach { (name:String) in
semaphoreSignal.wait()
LoadHelper.loadFile(name, queue) {
print(name + " load success....")
semaphoreSignal.signal()
}
}
let wirte = DispatchWorkItem(flags: .barrier) {
print("loading all success!!!")
}
queue.async(execute: wirte)
print("---------------")
RunLoop.current.run()
输出
loading 2.jpg in <NSThread: 0x1018396c0>{number = 2, name = (null)}
loading 1.jpg in <NSThread: 0x100f58390>{number = 3, name = (null)}
2.jpg load success....
1.jpg load success....
loading 3.jpg in <NSThread: 0x1018396c0>{number = 2, name = (null)}
loading 4.jpg in <NSThread: 0x100f58390>{number = 3, name = (null)}
4.jpg load success....
3.jpg load success....
loading 5.jpg in <NSThread: 0x1018396c0>{number = 2, name = (null)}
loading 6.jpg in <NSThread: 0x100f58390>{number = 3, name = (null)}
---------------
5.jpg load success....
6.jpg load success....
loading all success!!!
当然这也是阻塞的, 不过我们在项目开发中可以把它封装起来，最大并发数由外部传入
