冷流还是热流

https://www.youtube.com/watch?v=M8YtV47kaqA&t=607s

pl学习视频

什么是冷流？

 fun fibonacci(): Flow<BigInteger> = flow {var x = BigInteger.ZEROvar y = BigInteger.ONEwhile (true) {println("fibonacci while $x")emit(x)x = y.also {y += x}}}

这时候试着运行

    fun testJob() {testJob = viewModelScope.launch {repository.fibonacci()}

虽然运行了fibonacci()，但是没有任何的打印。原因是，冷流是只有有订阅者后，才开始发送数据。所以需要collect之后，才有数据

fun testJob() {testJob = viewModelScope.launch {var i = 0repository.fibonacci().take(100).collect {i++println("fibonacci $it")println("fibonacci $i")}}}

像之前的例子也有说到callbackflow，他也是一个冷流，之前有个例子，监听网络状态

fun networkAvailability(context: Context) = callbackFlow {val callback = object : ConnectivityManager.NetworkCallback() {override fun onAvailable(network: Network) {super.onAvailable(network)println("network send true")trySend(true)}override fun onLost(network: Network) {super.onLost(network)println("network send false")trySend(false)}}val manager = context.getSystemService(Context.CONNECTIVITY_SERVICE) as ConnectivityManager//注册println("network register")manager.registerNetworkCallback(NetworkRequest.Builder().addCapability(NetworkCapabilities.NET_CAPABILITY_INTERNET).build(), callback)awaitClose {//取消注册println("network unregister")manager.unregisterNetworkCallback(callback)}}

我们监听这个flow

 val flow = remember {         viewModel.networkAvailability(applicationContext)  }val status1 = flow.collectAsStateWithLifecycle(false)val status2 = flow.collectAsStateWithLifecycle(false)

看下打印情况

可以看到网络监听注册了两次，所以对冷流的数据收集是独立的。这不是我们想要的，我们希望只监听一次，然后只关心最新的数据，这个数据可以多个观察者共享，这时候，就希望把冷流变成热流，shareIn

                    val rememberCoroutineScope = rememberCoroutineScope()val flow = remember { viewModel.networkAvailability(applicationContext).shareIn(rememberCoroutineScope,SharingStarted.Eagerly)}val status1 = flow.collectAsStateWithLifecycle(false)val status2 = flow.collectAsStateWithLifecycle(false)

只是将flow转换 shareIn之后