启动多个异步任务并在它们完成时处理它们 (C#)

Question

所以我正在尝试学习如何编写异步方法，并且一直在努力让异步调用工作。似乎总是发生的是代码挂在 "await" 指令上，直到它最终似乎超时并以相同的方法使加载表单崩溃。

这很奇怪有两个主要原因：

当非异步且只是一个简单的循环时，代码可以完美运行
我几乎一字不差地复制了 MSDN 代码以将代码转换为此处的异步调用：https://msdn.microsoft.com/en-us/library/mt674889.aspx

我知道表格上已经有很多关于这个的问题，但我已经完成了其中的大部分，并尝试了很多其他方法（结果相同），现在似乎认为在之后出现了根本性的错误MSDN 代码无效。

后台工作者调用的主要方法如下：

// this method loads data from each individual webPage
async Task LoadSymbolData(DoWorkEventArgs _e)
{
    int MAX_THREADS = 10;
    int tskCntrTtl = dataGridView1.Rows.Count;
    Dictionary<string, string> newData_d = new Dictionary<string, string>(tskCntrTtl);
    // we need to make copies of things that can change in a different thread
    List<string> links = new List<string>(dataGridView1.Rows.Cast<DataGridViewRow>()
        .Select(r => r.Cells[dbIndexs_s.url].Value.ToString()).ToList());
    List<string> symbols = new List<string>(dataGridView1.Rows.Cast<DataGridViewRow>()
        .Select(r => r.Cells[dbIndexs_s.symbol].Value.ToString()).ToList());
    // we need to create a cancelation token once this is working
    // TODO

    using (LoadScreen loadScreen = new LoadScreen("Querying stock servers..."))
    {
        // we cant use the delegate becaus of async keywords
        this.loaderScreens.Add(loadScreen);
        // wait until the form is loaded so we dont get exceptions when writing to controls on that form
        while ( !loadScreen.IsLoaded() );
        // load the total number of operations so we can simplify incrementing the progress bar
        // on seperate form instances
        loadScreen.LoadProgressCntr(0, tskCntrTtl);
        // try to run all async tasks since they are non-blocking threaded operations
        for (int i = 0; i < tskCntrTtl; i += MAX_THREADS)
        {
            List<Task<string[]>> ProcessURL = new List<Task<string[]>>();
            List<int> taskList = new List<int>();

            // Make a list of task indexs
            for (int task = i; task < i + MAX_THREADS && task < tskCntrTtl; task++)
                taskList.Add(task);

            // ***Create a query that, when executed, returns a collection of tasks.
            IEnumerable<Task<string[]>> downloadTasksQuery =
                from task in taskList select QueryHtml(loadScreen, links[task], symbols[task]);

            // ***Use ToList to execute the query and start the tasks. 
            List<Task<string[]>> downloadTasks = downloadTasksQuery.ToList();

            // ***Add a loop to process the tasks one at a time until none remain.
            while (downloadTasks.Count > 0)
            {
                // Identify the first task that completes.
                Task<string[]> firstFinishedTask = await Task.WhenAny(downloadTasks);   // <---- CODE HANGS HERE

                // ***Remove the selected task from the list so that you don't
                // process it more than once.
                downloadTasks.Remove(firstFinishedTask);

                // Await the completed task.
                string[] data = await firstFinishedTask;
                if (!newData_d.ContainsKey(data.First())) 
                    newData_d.Add(data.First(), data.Last());
            }
        }
        // now we have the dictionary with all the information gathered from teh websites
        // now we can add the columns if they dont already exist and load the information
        // TODO
        loadScreen.UpdateProgress(100);
        this.loaderScreens.Remove(loadScreen);
    }
}

这里是查询网页的异步方法：

async Task<string[]> QueryHtml(LoadScreen _loadScreen, string _link, string _symbol) 
{
    string data = String.Empty;

    try
    {
        HttpClient client = new HttpClient();
        var doc = new HtmlAgilityPack.HtmlDocument();
        var html = await client.GetStringAsync(_link);    // <---- CODE HANGS HERE
        doc.LoadHtml(html);

        string percGrn = doc.FindInnerHtml(
            "//span[contains(@class,'time_rtq_content') and contains(@class,'up_g')]//span[2]");
        string percRed = doc.FindInnerHtml(
            "//span[contains(@class,'time_rtq_content') and contains(@class,'down_r')]//span[2]");

        // create somthing we'll nuderstand later
        if ((String.IsNullOrEmpty(percGrn) && String.IsNullOrEmpty(percRed)) ||
            (!String.IsNullOrEmpty(percGrn) && !String.IsNullOrEmpty(percRed)))
            throw new Exception();

        // adding string to empty gives string
        string perc = percGrn + percRed;
        bool isNegative = String.IsNullOrEmpty(percGrn);
        double percDouble;

        if (double.TryParse(Regex.Match(perc, @"\d+([.])?(\d+)?").Value, out percDouble))
            data = (isNegative ? 0 - percDouble : percDouble).ToString();
    }
    catch (Exception ex) { }
    finally
    {
        // update the progress bar...
        _loadScreen.IncProgressCntr();
    }

    return new string[] { _symbol, data };
}

我真的需要一些帮助。谢谢！

Answer 1

简而言之，当您将异步与任何 'regular' 任务函数结合使用时，您会遇到死锁

http://olitee.com/2015/01/c-async-await-common-deadlock-scenario/

解决方案是使用 configureawait

var html = await client.GetStringAsync(_link).ConfigureAwait(false);

你需要这个的原因是你没有等待你的原始线程。

// ***Create a query that, when executed, returns a collection of tasks.
IEnumerable<Task<string[]>> downloadTasksQuery = from task in taskList select QueryHtml(loadScreen,links[task], symbols[task]);

这里发生的事情是您将 await 范例与三个常规任务处理范例混合在一起。并且那些不混合（或者更确切地说，您必须使用 ConfigureAwait(false) 才能工作。

启动多个异步任务并在它们完成时处理它们 (C#)

Starting Multiple Async Tasks and Process Them As They Complete (C#)

.net

c#

asynchronous

html-agility-pack