0

我在 C# 中有一个 Windows Phone 7 (7.1) 方法，它给定一个字符串形式的 URL，将该 URL 的内容下载到一个文件中（参见下面的代码）。从代码中可以看出，我分配了一个DownloadProgressChanged()事件处理程序。在该处理程序中，如果调用者提供了IProgress对象，我将调用该对象的Report()方法。考虑到用户的网络连接速度可能很慢，我想确保下载速度尽可能快。在WebClient 的DownloadProgressChanged()回调中调用IProgress.Report()方法会大大减慢下载速度吗？

我对IProgress.Report()不够熟悉，无法知道它是在当前线程还是调用线程上执行。它是否在调用线程上执行？我担心的是重复的线程切换真的会让事情陷入困境。我可能会将对该方法的调用包装在Task.Run()调用中，以保持 UI 线程满意。但以防万一，我会问我的代码是否有任何潜在的问题，就陷入 UI 线程而言？

对与结构或性能有关的代码的任何其他评论表示赞赏。请注意，我在此应用中使用Microsoft.Bcl.Async包。

更新（稍后添加）：关于线程切换，显然在 UI 线程上引发了 DownloadProgressChanged() 事件，而不是在下载线程上，因此无需对 Dispatcher 等进行任何花哨的操作来进行 UI 更新设想。至少根据此代码项目文章：

C# 5 Async 中的进度报告

    public static void URLToFile(string strUrl, string strDestFilename, IProgress<int> progress, int iNumSecondsToWait = 30)
    {
        strUrl = strUrl.Trim();

        if (String.IsNullOrWhiteSpace(strUrl))
            throw new ArgumentException("(Misc::URLToFile) The URL is empty.");

        strDestFilename = strDestFilename.Trim();

        if (String.IsNullOrWhiteSpace(strDestFilename))
            throw new ArgumentException("(Misc::URLToFile) The destination file name is empty.");

        if (iNumSecondsToWait < 1)
            throw new ArgumentException("(Misc::URLToFile) The number of seconds to wait is less than 1.");

        // Create the isolated storage file.
        StreamWriter sw = openIsoStorFileAsStreamWriter(strDestFilename);

        // If the stream writer is NULL, then the file could not be created.
        if (sw == null)
            throw new System.IO.IOException("(Misc::URLToFile) Error creating or writing to the file named: " + strDestFilename);

        // Asynchronous download.  Note, the Silverlight version of WebClient does *not* implement 
        //  IDisposable.
        WebClient wc = new WebClient();

        try
        {
            // Create a download progress changed handler so we can pass on progress
            //  reports to the caller if they provided a progress report object.
            wc.DownloadProgressChanged += (s, e) =>
            {
                // Do we have a progress report handler?
                if (progress != null)
                    // Yes, call it.
                    progress.Report(e.ProgressPercentage);
            };

            // Use a Lambda expression for the "completed" handler
            //  that writes the contents to a file.
            wc.OpenReadCompleted += (s, e) =>
                e.Result.CopyTo(sw.BaseStream);

            // Now make the call to download the file.
            wc.DownloadStringAsync(new Uri(strUrl));
        }
        finally
        {
            // Make sure the stream is cleaned up.
            sw.Flush();
            sw.Close();

            // Make sure the StreamWriter is diposed of.
            sw.Dispose();
        } // try/finally

        // CancellationTokenSource srcCancelToken = new CancellationTokenSource();
        // srcCancelToken.CancelAfter(TimeSpan.FromSeconds(iNumSecondsToWait));
    } // public static void URLToFile()
4

1 回答 1

1

No, the download progress check does not necessarily affect the download speed. Mainly because the items that are checked are not downloaded per-se. When you initiate the download, you get a size declaration (content length) - that is used as a reference for a complete file. Then, the size of the local (downloaded) byte content is checked and a ratio can be built from the two values.

NOTE: Does not apply for streaming, for obvious reasons, since there is no final size estimate.

于 2013-04-05T03:59:27.640 回答