Question

我正在通过下面提供的代码学习文件的MD5哈希值。但是，随着文件大小的增加，计算也将花费很长时间。我想将此计算反映在进度条对象上，但我不知道。

我想要这样的东西；

progressBar.Value = mD5.ComputedBytes;
progressBar.Maximum = mD5.TotalBytesToCompute;

如何制作？

代码；

public static string getMD5HashFromFile(string fileName)
{
    string str = "";
    using (MD5 mD5 = MD5.Create())
    {
        using (FileStream fileStream = File.OpenRead(fileName))
        { str = BitConverter.ToString(mD5.ComputeHash(fileStream)).Replace("-", string.Empty); fileStream.Close(); }
    }
    return str;
}

Answer 1

HashAlgorithm使您能够使用TransformBlock和TransformFinalBlock方法对数据块进行哈希处理。另一方面，Stream类还允许您异步读取大块数据。

考虑到这些事实，您可以创建一种方法来获取流作为输入，然后以块的形式读取流，然后为每个卡盘散列它，并通过计算读取的字节来报告进度（字节进程数）。

ComputeHashAsync

在这里，我为ComputeHashAsync类创建了HashAlgorithm扩展方法。它接受：

stream：输入Stream以计算哈希。
cancellationToken：可选的CancellationToken，可用于取消操作
progress：IProgress<long>的可选实例，用于接收进度报告（已处理的字节数）。
buggerSize：用于读取数据的可选缓冲区大小。默认ID 1024 * 1024字节。

代码如下：

using System;
using System.IO;
using System.Security.Cryptography;
using System.Threading;
using System.Threading.Tasks;

public static class HashAlgorithmExtensions {
    public static async Task<byte[]> ComputeHashAsync(
        this HashAlgorithm hashAlgorithm, Stream stream,
        CancellationToken cancellationToken = default(CancellationToken),
        IProgress<long> progress = null,
        int bufferSize = 1024 * 1024) {
        byte[] readAheadBuffer, buffer, hash;
        int readAheadBytesRead, bytesRead;
        long size, totalBytesRead = 0;
        size = stream.Length;
        readAheadBuffer = new byte[bufferSize];
        readAheadBytesRead = await stream.ReadAsync(readAheadBuffer, 0, 
           readAheadBuffer.Length, cancellationToken);
        totalBytesRead += readAheadBytesRead;
        do {
            bytesRead = readAheadBytesRead;
            buffer = readAheadBuffer;
            readAheadBuffer = new byte[bufferSize];
            readAheadBytesRead = await stream.ReadAsync(readAheadBuffer, 0,
                readAheadBuffer.Length, cancellationToken);
            totalBytesRead += readAheadBytesRead;

            if (readAheadBytesRead == 0)
                hashAlgorithm.TransformFinalBlock(buffer, 0, bytesRead);
            else
                hashAlgorithm.TransformBlock(buffer, 0, bytesRead, buffer, 0);
            if (progress != null)
                progress.Report(totalBytesRead);
            if (cancellationToken.IsCancellationRequested)
                cancellationToken.ThrowIfCancellationRequested();
        } while (readAheadBytesRead != 0);
        return hash = hashAlgorithm.Hash;
    }
}

示例1-更新ProgressBar

byte[] bytes;
using (var hash = MD5.Create())
{
    using (var fs = new FileStream(f, FileMode.Open))
    {
        bytes = await hash.ComputeHashAsync(fs,
            progress: new Progress<long>(i =>
            {
                progressBar1.Invoke(new Action(() =>
                {
                    progressBar1.Value = i;
                }));
            }));
        MessageBox.Show(BitConverter.ToString(bytes).Replace("-", string.Empty));
    }
}

示例2-1秒后取消任务

try
{
    var s = new CancellationTokenSource();
    s.CancelAfter(1000);
    byte[] bytes;
    using (var hash = MD5.Create())
    {
        using (var fs = new FileStream(f, FileMode.Open))
        {
            bytes = await hash.ComputeHashAsync(fs,
                cancellationToken: s.Token,
                progress: new Progress<long>(i =>
                {
                    progressBar1.Invoke(new Action(() =>
                    {
                        progressBar1.Value = i;
                    }));
                }));

            MessageBox.Show(BitConverter.ToString(bytes).Replace("-", string.Empty));
        }
    }
}
catch (OperationCanceledException)
{
    MessageBox.Show("Operation canceled.");
}

创建大型文件进行测试

var f = Path.Combine(Application.StartupPath, "temp.log");
File.Delete(f);
using (var fs = new FileStream(f, FileMode.Create))
{
    fs.Seek(1L * 1024 * 1024 * 1024, SeekOrigin.Begin);
    fs.WriteByte(0);
    fs.Close();
}

_{注意：大块计算哈希的实现来自Alexandre Gomes的博客post，然后我更改了代码使其成为async，并支持{{1} }和CancellationToken。}

报告哈希进度

1 个答案: