我对使用async修饰符进行异步编程并不熟悉。我试图弄清楚如何确保我的控制台应用程序的主要方法实际上是异步运行的。
class Program
{
static void Main(string[] args)
{
Bootstrapper bs = new Bootstrapper();
var list = bs.GetList();
}
}
public class Bootstrapper {
public async Task<List<TvChannel>> GetList()
{
GetPrograms pro = new GetPrograms();
return await pro.DownloadTvChannels();
}
}
我知道这不是从“顶层”异步运行的。由于不可能在Main方法上指定异步修饰符，我如何在Main中异步运行代码?
当前回答
下面的代码可以用来创建一个主异步。我已经调整它使用长时间运行的任务(了解更多信息在这里:https://learn.microsoft.com/en-us/dotnet/api/system.threading.tasks.taskcreationoptions?view=net-7.0)
它还实现了来自上述响应的取消令牌。
private static int Main(string[] args)
{
var cts = new CancellationTokenSource();
Console.CancelKeyPress += (s, e) =>
{
e.Cancel = !cts.IsCancellationRequested;
cts.Cancel();
Console.WriteLine("CancellationRequested");
};
try
{
var task = new Task<int>(
() => MainAsync(args, cts.Token).GetAwaiter().GetResult(),
cts.Token,
TaskCreationOptions.LongRunning //https://learn.microsoft.com/en-us/dotnet/api/system.threading.tasks.taskcreationoptions?view=net-7.0
);
task.Start();
var exitCode = task.GetAwaiter().GetResult();
/*Or this.*/
//var exitCode = MainAsync(args, cts.Token).GetAwaiter().GetResult();
return exitCode;// MainAsync(args, cts.Token).GetAwaiter().GetResult();
}
catch (OperationCanceledException ex)
{
Console.WriteLine(ex);
return 1223; // Cancelled.
}
catch(Exception ex)
{
Console.WriteLine(ex);
return -1;
}
}
private static async Task<int> MainAsync(string[] args, CancellationToken cancellationToken)
{
await Something()
return;
}
在下面的例子中，我写了。你可以使用maxDegreeOfParallelism & numberOfIteration来理解/查看任务是如何处理的。学习TPL很好的起点!
private static async Task<int> MainAsync(string[] args, CancellationToken cancellationToken)
{
var infos = new ConcurrentBag<Info>();
var mySuperUselessService = new BigWorkload();
int numberOfSecond = 1;
int numberOfIteration = 25; //Experiment with this
int maxDegreeOfParallelism = 4; //Experiment with this
var simulateWorkTime = TimeSpan.FromSeconds(numberOfSecond);
var informations = Enumerable.Range(1, numberOfIteration)
.Select(x => new Info() { Index = x });
var count = informations.Count();
var chunkNeeded = Math.Round(count / Convert.ToDecimal(maxDegreeOfParallelism), MidpointRounding.ToPositiveInfinity);
var splashInfo = @$"
Press CTRL + C to cancel.
Processing {count} items, maxDegreeOfParallelism set to {maxDegreeOfParallelism}.
But it will be bound by the core on the machine {Environment.ProcessorCount}.
This operation should take ~{chunkNeeded * (numberOfSecond + 0.01m)}s
And will be starting test in 2s
";
Console.WriteLine(splashInfo);
await Task.Delay(TimeSpan.FromSeconds(2));
var parralelOptions = new ParallelOptions() { MaxDegreeOfParallelism = maxDegreeOfParallelism, CancellationToken = cancellationToken};
var stopwatch = new Stopwatch();
stopwatch.Start();
var forLoopTask = Parallel.ForEachAsync(informations, parralelOptions, async (info, token) =>
{
await mySuperUselessService.Simulate(simulateWorkTime, info);
Console.WriteLine(info);
infos.Add(info);
});
await forLoopTask;
stopwatch.Stop();
foreach (var grouped in infos.GroupBy(x => x.ManagedThreadId))
{
Console.WriteLine($"ThreadId: {grouped.Key}");
foreach (var item in grouped)
{
Console.WriteLine($"\t Index: {item.Index} {item.TaskCurrentId}");
}
}
Console.WriteLine($"NumberOfThread: {infos.GroupBy(x => x.ManagedThreadId).Count()}");
Console.WriteLine($"Elasped: {stopwatch.ElapsedMilliseconds / 1000d}s");
Console.WriteLine(splashInfo);
return 0;
}
其他回答
当c# 5 CTP被引入时，你当然可以用async…尽管这样做通常不是个好主意。我相信这是在VS 2013的发布中改变的，成为一个错误。
除非你已经启动了任何其他前台线程，否则你的程序将在Main完成时退出，即使它已经启动了一些后台工作。
你到底想做什么?请注意，你的GetList()方法目前真的不需要是异步的——它添加了一个额外的层，没有真正的原因。它在逻辑上等价于(但比):
public Task<List<TvChannel>> GetList()
{
return new GetPrograms().DownloadTvChannels();
}
你可以用这个简单的构造来解决这个问题:
class Program
{
static void Main(string[] args)
{
Task.Run(async () =>
{
// Do any async anything you need here without worry
}).GetAwaiter().GetResult();
}
}
这将把你所做的所有事情放在你想要它的线程池(这样你启动/等待的其他任务就不会试图重新加入一个不应该的线程)，并等待所有事情都完成后再关闭控制台应用程序。不需要特殊的循环或外部库。
编辑:合并Andrew的未捕获异常的解决方案。
在MSDN上，Task的文档。Run Method (Action)提供了这个例子，展示了如何从main异步运行一个方法:
using System;
using System.Threading;
using System.Threading.Tasks;
public class Example
{
public static void Main()
{
ShowThreadInfo("Application");
var t = Task.Run(() => ShowThreadInfo("Task") );
t.Wait();
}
static void ShowThreadInfo(String s)
{
Console.WriteLine("{0} Thread ID: {1}",
s, Thread.CurrentThread.ManagedThreadId);
}
}
// The example displays the following output:
// Application thread ID: 1
// Task thread ID: 3
请注意下面的例子:
这些例子表明异步任务执行在不同的 线程比主应用程序线程多。
因此，如果您希望任务在主应用程序线程上运行，请参阅@StephenCleary的回答。
关于任务运行的线程，也请注意Stephen对他的回答的评论:
您可以使用简单的Wait或Result，没有任何问题 的。但要注意两者之间有两个重要的区别: 所有异步延续都运行在线程池上，而不是主线程池上 2)任何异常都包装在AggregateException中。
(参见异常处理(任务并行库)了解如何合并异常处理来处理AggregateException。)
最后，从MSDN上的任务文档。延迟方法(TimeSpan)，这个例子展示了如何运行一个返回值的异步任务:
using System;
using System.Threading.Tasks;
public class Example
{
public static void Main()
{
var t = Task.Run(async delegate
{
await Task.Delay(TimeSpan.FromSeconds(1.5));
return 42;
});
t.Wait();
Console.WriteLine("Task t Status: {0}, Result: {1}",
t.Status, t.Result);
}
}
// The example displays the following output:
// Task t Status: RanToCompletion, Result: 42
注意，不是传递一个委托给Task。运行时，你可以像这样传递一个lambda函数:
var t = Task.Run(async () =>
{
await Task.Delay(TimeSpan.FromSeconds(1.5));
return 42;
});
在Main中尝试将GetList调用更改为:
Task.Run(() => bs.GetList());
下面的代码可以用来创建一个主异步。我已经调整它使用长时间运行的任务(了解更多信息在这里:https://learn.microsoft.com/en-us/dotnet/api/system.threading.tasks.taskcreationoptions?view=net-7.0)
它还实现了来自上述响应的取消令牌。
private static int Main(string[] args)
{
var cts = new CancellationTokenSource();
Console.CancelKeyPress += (s, e) =>
{
e.Cancel = !cts.IsCancellationRequested;
cts.Cancel();
Console.WriteLine("CancellationRequested");
};
try
{
var task = new Task<int>(
() => MainAsync(args, cts.Token).GetAwaiter().GetResult(),
cts.Token,
TaskCreationOptions.LongRunning //https://learn.microsoft.com/en-us/dotnet/api/system.threading.tasks.taskcreationoptions?view=net-7.0
);
task.Start();
var exitCode = task.GetAwaiter().GetResult();
/*Or this.*/
//var exitCode = MainAsync(args, cts.Token).GetAwaiter().GetResult();
return exitCode;// MainAsync(args, cts.Token).GetAwaiter().GetResult();
}
catch (OperationCanceledException ex)
{
Console.WriteLine(ex);
return 1223; // Cancelled.
}
catch(Exception ex)
{
Console.WriteLine(ex);
return -1;
}
}
private static async Task<int> MainAsync(string[] args, CancellationToken cancellationToken)
{
await Something()
return;
}
在下面的例子中，我写了。你可以使用maxDegreeOfParallelism & numberOfIteration来理解/查看任务是如何处理的。学习TPL很好的起点!
private static async Task<int> MainAsync(string[] args, CancellationToken cancellationToken)
{
var infos = new ConcurrentBag<Info>();
var mySuperUselessService = new BigWorkload();
int numberOfSecond = 1;
int numberOfIteration = 25; //Experiment with this
int maxDegreeOfParallelism = 4; //Experiment with this
var simulateWorkTime = TimeSpan.FromSeconds(numberOfSecond);
var informations = Enumerable.Range(1, numberOfIteration)
.Select(x => new Info() { Index = x });
var count = informations.Count();
var chunkNeeded = Math.Round(count / Convert.ToDecimal(maxDegreeOfParallelism), MidpointRounding.ToPositiveInfinity);
var splashInfo = @$"
Press CTRL + C to cancel.
Processing {count} items, maxDegreeOfParallelism set to {maxDegreeOfParallelism}.
But it will be bound by the core on the machine {Environment.ProcessorCount}.
This operation should take ~{chunkNeeded * (numberOfSecond + 0.01m)}s
And will be starting test in 2s
";
Console.WriteLine(splashInfo);
await Task.Delay(TimeSpan.FromSeconds(2));
var parralelOptions = new ParallelOptions() { MaxDegreeOfParallelism = maxDegreeOfParallelism, CancellationToken = cancellationToken};
var stopwatch = new Stopwatch();
stopwatch.Start();
var forLoopTask = Parallel.ForEachAsync(informations, parralelOptions, async (info, token) =>
{
await mySuperUselessService.Simulate(simulateWorkTime, info);
Console.WriteLine(info);
infos.Add(info);
});
await forLoopTask;
stopwatch.Stop();
foreach (var grouped in infos.GroupBy(x => x.ManagedThreadId))
{
Console.WriteLine($"ThreadId: {grouped.Key}");
foreach (var item in grouped)
{
Console.WriteLine($"\t Index: {item.Index} {item.TaskCurrentId}");
}
}
Console.WriteLine($"NumberOfThread: {infos.GroupBy(x => x.ManagedThreadId).Count()}");
Console.WriteLine($"Elasped: {stopwatch.ElapsedMilliseconds / 1000d}s");
Console.WriteLine(splashInfo);
return 0;
}
