Downloading files not affecting UI - c#

We are currently downloading files using HttpClient because our backend requires certificate.
I have a control - FileRowwhich is an UI element with some code-behind methods for file downloading, like this one:
if (FileIsDownloaded == false)
{
await CoreApplication.MainView.CoreWindow.Dispatcher.RunAsync(Windows.UI.Core.CoreDispatcherPriority.Low, () =>
{
DataManager.Instance.DownloadFile(this);
});
}
if (ThumbnailIsDownloaded == false)
{
await CoreApplication.MainView.CoreWindow.Dispatcher.RunAsync(Windows.UI.Core.CoreDispatcherPriority.Low, () =>
{
DataManager.Instance.DownloadThumbnail(this);
});
}
Downloading single item is fine but when i click download all ( about 50 items ) the whole UI starts to freeze.
As you can see, i have tried to give requests low priority - but still same result.
Answers to common questions:
1) Yes the files should be downloadable all at one time, not one after another.
2) DataManager.Instance.DownloadThumbnail(this) - i do this to give refference to current control so that i could report a progress in a progress bar.
Any suggestions?
EDIT:
Downloading looks like this:
public async void DownloadFile(FileRow fileRow)
{
//Lot of checking for if file exist, if version is the same
string LocalFilename = await DownloadManager.DownloadFile(fileRow.MyFile.file.id, fileRow.MyFile.file.version, fileRow.MyFile.file.filename,fileRow);
// next is just using the filename string
}
And finally my download:
public static async Task<string> DownloadFileOfCustomerAssetRow(int? id, int? version, string filename, FileRow fileRow)
{
try
{
HttpClientHandler aHandler = new HttpClientHandler();
aHandler.ClientCertificateOptions = ClientCertificateOption.Automatic;
HttpClient aClient = new HttpClient(aHandler);
customerAssetRow.CurrentFileDownload = aClient;
aClient.DefaultRequestHeaders.ExpectContinue = false;
HttpResponseMessage response = await aClient.GetAsync(WebServices.BackendStartUrl + "getFileData?id=" + id + "&version=" + version, HttpCompletionOption.ResponseHeadersRead);
var file = await ApplicationData.Current.LocalFolder.CreateFileAsync(filename, Windows.Storage.CreationCollisionOption.GenerateUniqueName);
fileRow.FileName = file.Name;
using (var fs = await file.OpenAsync(Windows.Storage.FileAccessMode.ReadWrite))
{
Stream stream = await response.Content.ReadAsStreamAsync();
IInputStream inputStream = stream.AsInputStream();
ulong totalBytesRead = 0;
while (true)
{
IBuffer buffer = new Windows.Storage.Streams.Buffer(1024);
buffer = await inputStream.ReadAsync(
buffer,
buffer.Capacity,
InputStreamOptions.None);
if (buffer.Length == 0)
{
break;
}
totalBytesRead += buffer.Length;
fileRow.Progress.Value = fileRow.Progress.Value + 1024;
await fs.WriteAsync(buffer);
}
inputStream.Dispose();
}
fileRow.Progress.Visibility = Visibility.Collapsed;
return file.Name;
}
catch (Exception e)
{
ErrorReporter.ReportError("Error in DownloadManager.cs in function DownloadFile.", e);
return "";
}
}

It's possible that your async method continuations are interrupting the UI thread too much. Try creating a stronger separation between your background logic (DownloadFileOfCustomerAssetRow) and your UI (FileRow) by introducing an IProgress<T> reporter. Then ensure that every await in your background logic has a ConfigureAwait(false) on it.
public static async Task<string> DownloadFileOfCustomerAssetRow(int? id, int? version, string filename, IProgress<int> progress)
{
HttpClientHandler aHandler = new HttpClientHandler();
aHandler.ClientCertificateOptions = ClientCertificateOption.Automatic;
HttpClient aClient = new HttpClient(aHandler);
customerAssetRow.CurrentFileDownload = aClient;
aClient.DefaultRequestHeaders.ExpectContinue = false;
HttpResponseMessage response = await aClient.GetAsync(WebServices.BackendStartUrl + "getFileData?id=" + id + "&version=" + version, HttpCompletionOption.ResponseHeadersRead).ConfigureAwait(false);
var file = await ApplicationData.Current.LocalFolder.CreateFileAsync(filename, Windows.Storage.CreationCollisionOption.GenerateUniqueName).ConfigureAwait(false);
fileRow.FileName = file.Name;
using (var fs = await file.OpenAsync(Windows.Storage.FileAccessMode.ReadWrite).ConfigureAwait(false))
{
Stream stream = await response.Content.ReadAsStreamAsync().ConfigureAwait(false);
IInputStream inputStream = stream.AsInputStream();
ulong totalBytesRead = 0;
while (true)
{
IBuffer buffer = new Windows.Storage.Streams.Buffer(1024);
buffer = await inputStream.ReadAsync(
buffer,
buffer.Capacity,
InputStreamOptions.None).ConfigureAwait(false);
if (buffer.Length == 0)
{
break;
}
totalBytesRead += buffer.Length;
if (progress != null)
progress.Report(totalBytesRead);
await fs.WriteAsync(buffer).ConfigureAwait(false);
}
inputStream.Dispose();
}
return file.Name;
}

Related

Get the value of download progress when using HttpClient in WPF [duplicate]

This question already has answers here:
Stream.CopyToAsync with progress reporting - progress is reported even after copying finish
(2 answers)
Progress bar with HttpClient
(11 answers)
Closed last month.
How would you get the current progress as a interger for example when downloading a file?
I've been trying to implementing it into the code below with no luck at all
public static async Task HttpDownload(string url, string OutputLocation)
{
using (HttpClient client = new HttpClient())
using (var response = await client.GetAsync(url, HttpCompletionOption.ResponseHeadersRead))
{
var contentLength = response.Content.Headers.ContentLength;
using (Stream streamToReadFrom = await response.Content.ReadAsStreamAsync())
{
string fileToWriteTo = OutputLocation;
using (Stream streamToWriteTo = File.Open(fileToWriteTo, FileMode.Create))
{
await streamToReadFrom.CopyToAsync(streamToWriteTo);
}
}
}
}
I've tried things like Progress<T>, but it i could never covert it to a usage value
(Edit, I found something thats actually works for me, Progress bar with HttpClient)
I wrote my own method to achieve this. Its complettely stream based. So almost no load on your system memory. But feel free to test it out.
public static async Task<T> DownloadAsync<T>(
this HttpClient client,
Uri requestUri,
Func<HttpResponseMessage, T> destinationStreamFactory,
IProgress<HttpDownloadProgress> progress = null,
CancellationToken cancellationToken = default)
where T : Stream
{
// Get the http headers first to examine the content length
using (var response = await client.GetAsync(requestUri, HttpCompletionOption.ResponseHeadersRead, cancellationToken))
{
var contentLength = response.Content.Headers.ContentLength;
using (var download = await response.Content.ReadAsStreamAsync())
{
var destinationStream = destinationStreamFactory(response);
// Ignore progress reporting when no progress reporter was
// passed or when the content length is unknown
if (progress == null)
{
await download.CopyToAsync(destinationStream, 81920, cancellationToken);
return destinationStream;
}
// Convert absolute progress (bytes downloaded) into relative progress (0% - 100%)
var relativeProgress = new Progress<long>(totalBytes => progress.Report(
new HttpDownloadProgress()
{
ContentLength = contentLength ?? 0,
RelativeProgress = contentLength.HasValue ? Convert.ToInt32(100 * (float)totalBytes / contentLength.Value) : 0,
ReadBytes = totalBytes
}));
// Use extension method to report progress while downloading
await download.CopyToAsync(destinationStream, 81920, relativeProgress, cancellationToken);
progress.Report(new HttpDownloadProgress()
{
ContentLength = contentLength ?? 0,
RelativeProgress = 1,
ReadBytes = contentLength ?? 0
});
return destinationStream;
}
}
Example:
var downloadProgress = new Progress<HttpDownloadProgress>(
f =>
{
var readMbs = f.ReadBytes / 1024 / 1024;
var contentMbs = f.ContentLength / 1024 / 1024;
Progress = f.RelativeProgress;
Message = $"Downloading {pluginVersion.PluginName} ({readMbs}MB/{contentMbs}MB)";
});
using (var fileStream = await client.DownloadAsync(
serviceEndPoint,
response => CreateFileStream(response, outputPath),
downloadProgress,
token))
{
Console.WriteLine($"File name {fileStream.Name}";
}
with the FileStreamFactory, which creates the output folder and a filestream with the extracted file name of the httpreponse:
private FileStream CreateFileStream(HttpResponseMessage response, string outputPath)
{
// get the actual content stream
var fileName = response.Content.Headers.ContentDisposition?.FileName;
if (fileName == null)
{
throw new InvalidOperationException("No filename available");
}
var outputFile = Path.Combine(outputPath, fileName);
var folder = Directory.GetParent(outputFile)?.FullName;
if (folder != null && !Directory.Exists(folder))
{
Directory.CreateDirectory(folder);
}
else if (File.Exists(outputFile))
{
File.Delete(outputFile);
}
Console.WriteLine($"Downloading file to '{outputFile} ...");
return File.Open(outputFile, FileMode.Create);
}
public class HttpDownloadProgress
{
public int RelativeProgress { get; set; }
public long ContentLength { get; set; }
public long ReadBytes { get; set; }
}

HTTPS proxy implementation, how to detect a completed request

I'm attempting to write a simple async https proxy server in c#.
I would like to know how I should detect/handle when the request is complete, and how to exit my bActive loop, assuming a loop like this is appropriate.
Would really appreciate some pointers on if my approach is correct and what I could do to improve the logic.
The issue I seem to be running into is that the time it takes for an endpoint to respond along with the network delay means I DataAvailable doenst always have data but there may still be some sending. Requiring a sleep and another attmempt which in turn causes the long completion time in requests.
Listen for TCP connection
Extract CONNECT header and open a connection to the requested server
Copy the requestStream to proxyStream
Copy the proxyStream to the requestStream
Sleep waiting for data and repeat 3 - 4 until no data is avaiable on both streams. Then break out of the loop and close connection.
public async Task Start()
{
listener.Start();
while (listen)
{
if (listener.Pending())
{
HandleClient(await listener.AcceptTcpClientAsync());
}
else
{
await Task.Delay(100); //<--- timeout
}
}
}
private static async Task HandleClient(TcpClient clt)
{
var bytes = new byte[clt.ReceiveBufferSize];
var hostHeaderAvailable = 0;
NetworkStream requestStream = null;
int count;
const string connectText = "connect";
const string hostText = "Host: ";
bool bActive = true;
List<Task> tasks = new List<Task>();
try
{
using (NetworkStream proxyStream = clt.GetStream())
using (TcpClient requestClient = new TcpClient())
{
proxyStream.ReadTimeout = 100;
proxyStream.WriteTimeout = 100;
while (bActive)
{
if (proxyStream.DataAvailable && hostHeaderAvailable == 0)
{
count = await proxyStream.ReadAsync(bytes, 0, bytes.Length);
var text = Encoding.UTF8.GetString(bytes);
Console.WriteLine(text);
if (text.ToLower().StartsWith(connectText))
{
// extract the url and port
var host = text.Remove(0, connectText.Length + 1);
var hostIndex = host.IndexOf(" ", StringComparison.Ordinal);
var hostEntry = host.Remove(hostIndex).Split(new[] { ":" }, StringSplitOptions.None);
// connect to the url and prot supplied
await requestClient.ConnectAsync(hostEntry[0], Convert.ToInt32(hostEntry[1]));
requestStream = requestClient.GetStream();
requestStream.ReadTimeout = 100;
requestStream.WriteTimeout = 100;
// send 200 response to proxyStream
const string sslResponse = "HTTP/1.0 200 Connection established\r\n\r\n";
var sslResponseBytes = Encoding.UTF8.GetBytes(sslResponse);
await proxyStream.WriteAsync(sslResponseBytes, 0, sslResponseBytes.Length);
// delay here seems to prevent the following proxyStream.read from failing as data is not yet avaiable
// without it the loop runs and has to timeout before running again
await Task.Delay(1);
}
}
hostHeaderAvailable++;
if (requestStream == null || !requestClient.Connected || !clt.Connected)
{
bActive = false;
break;
}
Console.WriteLine(proxyStream.DataAvailable || requestStream.DataAvailable);
if (proxyStream.DataAvailable || requestStream.DataAvailable)
{
Task task = proxyStream.CopyToAsync(requestStream);
Task task2 = requestStream.CopyToAsync(proxyStream);
tasks.Add(task);
tasks.Add(task2);
await Task.WhenAll(tasks).ConfigureAwait(false);
bActive = false;
break;
}
await Task.Delay(10);
}
}
}
catch (Exception e)
{
Console.WriteLine(e.ToString());
}
clt.Close();
}
An older attempt that used ReadAsync/WriteAsync too longer to response and still had the timeout issue.
Listen for TCP connection
Extract CONNECT header and open a connection to the requested server
Read data from requestStream and copy to proxyStream
Wait checking if data is avaiable on either stream
If data avaiable read from proxyStream and write to requestStream
If data avaiable read from requestStream and write to proxyStream
Sleep waiting for data and repeat 5 - 6 until no data is avaiable on eitboth streams. Then break out of the loop and close connection.
private static TcpListener listener = new TcpListener(IPAddress.Parse("192.168.0.25"), 13000);
private static bool listen = true;
public async Task Start()
{
listener.Start();
while (listen)
{
if (listener.Pending())
{
await HandleClient(await listener.AcceptTcpClientAsync());
}
else
{
await Task.Delay(100);
}
}
}
private static async Task HandleClient(TcpClient clt)
{
var bytes = new byte[clt.ReceiveBufferSize];
var hostHeaderAvailable = 0;
NetworkStream requestStream = null;
int count;
const string connectText = "connect";
const string hostText = "Host: ";
bool bActive = true;
try
{
using (NetworkStream proxyStream = clt.GetStream())
using (TcpClient requestClient = new TcpClient())
{
while (bActive)
{
while (proxyStream.DataAvailable)
{
// handle connect
if (hostHeaderAvailable == 0)
{
count = await proxyStream.ReadAsync(bytes, 0, bytes.Length);
var text = Encoding.UTF8.GetString(bytes);
Console.WriteLine(text);
if (text.ToLower().StartsWith(connectText))
{
// extract the url and port
var host = text.Remove(0, connectText.Length + 1);
var hostIndex = host.IndexOf(" ", StringComparison.Ordinal);
var hostEntry = host.Remove(hostIndex).Split(new[] { ":" }, StringSplitOptions.None);
// connect to the url and prot supplied
await requestClient.ConnectAsync(hostEntry[0], Convert.ToInt32(hostEntry[1]));
requestStream = requestClient.GetStream();
// send 200 response to proxyStream
const string sslResponse = "HTTP/1.0 200 Connection established\r\n\r\n";
var sslResponseBytes = Encoding.UTF8.GetBytes(sslResponse);
await proxyStream.WriteAsync(sslResponseBytes, 0, sslResponseBytes.Length);
// delay here seems to prevent the following proxyStream.read from failing as data is not yet avaiable
// without it the loop runs and has to timeout before running again
await Task.Delay(20);
}
}
hostHeaderAvailable++;
if (requestClient.Connected && hostHeaderAvailable > 1)
{
count = await proxyStream.ReadAsync(bytes, 0, bytes.Length);
await requestStream.WriteAsync(bytes, 0, count);
}
}
while (requestStream.DataAvailable)
{
count = await requestStream.ReadAsync(bytes, 0, bytes.Length);
await proxyStream.WriteAsync(bytes, 0, count);
}
// attempt to detect a timeout / end of data avaiable
var timeout = 0;
while (!proxyStream.DataAvailable && !requestStream.DataAvailable)
{
if (timeout > 5)
{
bActive = false;
break;
}
await Task.Delay(10);
timeout++;
}
}
}
}
catch (Exception e)
{
Console.WriteLine(e.ToString());
}
}
UPDATE
As per AgentFire's answer I have now come to the following working code:
public static async Task HandleDisconnect(TcpClient tcp, TcpClient tcp2, CancellationToken cancellationToken)
{
while (true)
{
if (tcp.Client.Poll(0, SelectMode.SelectRead))
{
byte[] buff = new byte[1];
if (tcp.Client.Receive(buff, SocketFlags.Peek) == 0)
{
// Client disconnected
Console.WriteLine("The requesting client has dropped its connection.");
cancellationToken = new CancellationToken(true);
break;
}
}
if (tcp2.Client.Poll(0, SelectMode.SelectRead))
{
byte[] buff = new byte[1];
if (tcp2.Client.Receive(buff, SocketFlags.Peek) == 0)
{
// Server disconnected
Console.WriteLine("The destination client has dropped its connection.");
cancellationToken = new CancellationToken(true);
break;
}
}
await Task.Delay(1);
}
}
private static async Task HandleClient(TcpClient clt)
{
List<Task> tasks = new List<Task>();
var bytes = new byte[clt.ReceiveBufferSize];
var hostHeaderAvailable = 0;
NetworkStream requestStream = null;
const string connectText = "connect";
try
{
using (NetworkStream proxyStream = clt.GetStream())
using (TcpClient requestClient = new TcpClient())
{
proxyStream.ReadTimeout = 100;
proxyStream.WriteTimeout = 100;
if (proxyStream.DataAvailable && hostHeaderAvailable == 0)
{
await proxyStream.ReadAsync(bytes, 0, bytes.Length);
var text = Encoding.UTF8.GetString(bytes);
Console.WriteLine(text);
if (text.ToLower().StartsWith(connectText))
{
// extract the url and port
var host = text.Remove(0, connectText.Length + 1);
var hostIndex = host.IndexOf(" ", StringComparison.Ordinal);
var hostEntry = host.Remove(hostIndex).Split(new[] { ":" }, StringSplitOptions.None);
// connect to the url and prot supplied
await requestClient.ConnectAsync(hostEntry[0], Convert.ToInt32(hostEntry[1]));
requestStream = requestClient.GetStream();
requestStream.ReadTimeout = 100;
requestStream.WriteTimeout = 100;
// send 200 response to proxyStream
const string sslResponse = "HTTP/1.0 200 Connection established\r\n\r\n";
var sslResponseBytes = Encoding.UTF8.GetBytes(sslResponse);
await proxyStream.WriteAsync(sslResponseBytes, 0, sslResponseBytes.Length);
}
}
hostHeaderAvailable++;
CancellationToken cancellationToken = new CancellationToken(false);
Task task = proxyStream.CopyToAsync(requestStream, cancellationToken);
Task task2 = requestStream.CopyToAsync(proxyStream, cancellationToken);
Task handleConnection = HandleDisconnect(clt, requestClient, cancellationToken);
tasks.Add(task);
tasks.Add(task2);
tasks.Add(handleConnection);
await Task.WhenAll(tasks).ConfigureAwait(false);
// close conenctions
clt.Close();
clt.Dispose();
requestClient.Close();
requestClient.Dispose();
}
}
catch (Exception e)
{
Console.WriteLine(e.ToString());
}
}
UPDATE
Attempt at using CancellationTokenSource
CancellationTokenSource source = new CancellationTokenSource();
CancellationToken cancellationToken = source.Token;
TaskFactory factory = new TaskFactory(cancellationToken);
tasks.Add(factory.StartNew(() => {proxyStream.CopyToAsync(requestStream);}, cancellationToken));
tasks.Add(factory.StartNew(() => {requestStream.CopyToAsync(proxyStream);}, cancellationToken));
tasks.Add(factory.StartNew(async () => {
//wait for this to retur, then cancel the token
await HandleDisconnect(clt, requestClient);
source.Cancel();
}, cancellationToken));
try
{
await factory.ContinueWhenAll(tasks.ToArray(),
(results) =>
{
Console.WriteLine("Tasks complete");
}, cancellationToken);
}
catch (AggregateException ae)
{
foreach (Exception e in ae.InnerExceptions)
{
if (e is TaskCanceledException)
Console.WriteLine("Unable to compute mean: {0}",
((TaskCanceledException)e).Message);
else
Console.WriteLine("Exception: " + e.GetType().Name);
}
}
finally
{
source.Dispose();
}
UPDATE
public static class extensionTcpClient{
public static bool CheckIfDisconnected(this TcpClient tcp)
{
if (tcp.Client.Poll(0, SelectMode.SelectRead))
{
byte[] buff = new byte[1];
if (tcp.Client.Receive(buff, SocketFlags.Peek) == 0)
{
// Client disconnected
return false;
}
}
return true;
}
}
class ProxyMaintainer
{
private static TcpListener listener = new TcpListener(IPAddress.Parse("192.168.0.25"), 13000);
public ProxyMaintainer()
{
}
public async Task Start()
{
Console.WriteLine("###############################");
Console.WriteLine("Listening on 192.168.0.25:13000");
Console.WriteLine("###############################\n");
listener.Start();
while (listen)
{
if (listener.Pending())
{
HandleClient(await listener.AcceptTcpClientAsync());
}
else
{
await Task.Delay(100); //<--- timeout
}
}
}
private static async Task Transport(NetworkStream from, NetworkStream to, Func<bool> isAlivePoller, CancellationToken token)
{
byte[] buffer = new byte[4096];
while (isAlivePoller())
{
while (from.DataAvailable)
{
int read = await from.ReadAsync(buffer, 0, buffer.Length, token).ConfigureAwait(false);
await to.WriteAsync(buffer, 0, read, token);
}
// Relieve the CPU a bit.
await Task.Delay(10, token).ConfigureAwait(false);
}
}
private static async Task HandleClient(TcpClient clientFrom)
{
var hostHeaderAvailable = 0;
int count;
var bytes = new byte[clientFrom.ReceiveBufferSize];
const string connectText = "connect";
NetworkStream toStream = null;
using (var fromStream = clientFrom.GetStream())
using(TcpClient clientTo = new TcpClient())
using (var manualStopper = new CancellationTokenSource())
{
count = await fromStream.ReadAsync(bytes, 0, bytes.Length);
var text = Encoding.UTF8.GetString(bytes);
Console.WriteLine(text);
if (text.ToLower().StartsWith(connectText))
{
// extract the url and port
var host = text.Remove(0, connectText.Length + 1);
var hostIndex = host.IndexOf(" ", StringComparison.Ordinal);
var hostEntry = host.Remove(hostIndex).Split(new[] { ":" }, StringSplitOptions.None);
// connect to the url and prot supplied
await clientTo.ConnectAsync(hostEntry[0], Convert.ToInt32(hostEntry[1]));
toStream = clientTo.GetStream();
// send 200 response to proxyStream
const string sslResponse = "HTTP/1.0 200 Connection established\r\n\r\n";
var sslResponseBytes = Encoding.UTF8.GetBytes(sslResponse);
await fromStream.WriteAsync(sslResponseBytes, 0, sslResponseBytes.Length);
}
bool Poller() => clientFrom.CheckIfDisconnected() && clientTo.CheckIfDisconnected();
Task one = Transport(fromStream, toStream, Poller, manualStopper.Token);
Task two = Transport(toStream, fromStream, Poller, manualStopper.Token);
await Task.WhenAll(one, two).ConfigureAwait(false);
//await one; await two; // To get exceptions if you want them and there are any.
// Alternatively, you can use Task.WhenAll to get exceptions aggregated for you.
}
Console.WriteLine("Closing connection");
}
}
Well, tell you what. The data availability, when it comes to HTTP, lies only in one parameter (if we omit things like WebSocket), which is called Connection and is passed as a Header as a one of two possible states: Close or Keep-Alive.
If Close is chosen by the client, the server is obliged to close the conection as soon as the request is served, whereas Keep-Alive tells the server that, if it doesn't want to, it may leave connection open for another request.
Let's consider both cases.
If client chooses Keep-Alive, the connection will persist and work as intended, indefinetely. But:
If either side drops the connection, there is an easy way to detect that. This piece of code was found on StackOverflow and it was told that it still works perfectly:
public static bool CheckIfDisconnected(this TcpClient tcp)
{
if (tcp.Client.Poll(0, SelectMode.SelectRead))
{
byte[] buff = new byte[1];
if (tcp.Client.Receive(buff, SocketFlags.Peek) == 0)
{
// Client disconnected
return true;
}
}
return false;
}
So I believe that you, as a proxy-server, are not obliged to manage connection states at all and can leave it to the actual communication parties. All you have to do is to detect when either of your connections - proxy or request - is dropped, drop the other one and call it a day.
P.S. Now, you also asked about asynchronicity.
I must add that TCP connections are considered full-duplex. which means you are free to create two async-running tasks, both reading and writing to their own sinks. My thoughts, it would be the optimal course of action.
To answer your other question
You are still using Stream.CopyToAsync which, as I have told you, is not going to succeed as long as any communicating party decides to wait a bit before sending another chunk of data.
You are also somewhat overcomplicating your solution.
I would put it this way:
async Task Transport(NetworkStream from, NetworkStream to, Func<bool> isAlivePoller, CancellationToken token)
{
byte[] buffer = new byte[4096];
while (isAlivePoller())
{
while (from.DataAvailable)
{
int read = await from.ReadAsync(buffer, 0, buffer.Length, token).ConfigureAwait(false);
await to.WriteAsync(buffer, 0, read, token).ConfigureAwait(false);
}
// Relieve the CPU a bit.
await Task.Delay(100, token).ConfigureAwait(false);
}
}
And then in your main code:
using TcpClient clientFrom = ...;
using TcpClient clientTo = ...;
using var fromStream = clientFrom.GetStream();
using var toStream = clientTo.GetStream();
using var manualStopper = new CancellationTokenSource();
bool Poller() => clientFrom.CheckIfDisconnected() && clientTo.CheckIfDisconnected();
Task one = Transport(fromStream, toStream, Poller, stopper.Token);
Task two = Transport(toStream, fromStream, Poller, stopper.Token);
await Task.WhenAny(one, two).ConfigureAwait(false);
//await one; await two; // To get exceptions if you want them and there are any.
// Alternatively, you can use Task.WhenAll to get exceptions aggregated for you.
And you are pretty much done here.

Web API service hangs on reading the stream

Description: I am modifying the ASP.NET Core Web API service (hosted in Windows Service) that supports resumable file uploads. This works fine and resumes file uploads in many failure conditions except one described below.
Problem: When the service is on ther other computer and the client is on mine and I unplug the cable on my computer, the client detects the absence of network while the service hangs on fileSection.FileStream.Read(). Sometimes the service detects the failure in 8 min, sometimes in 20, sometimes never.
I also noticed that after I unplug cable and stop the client, the service becomes stuck at Read() function and the file size is x KB, but when the service finally detects the exception some time later, it writes additional 4 KB to the file. This is weird because I turned off buffering and the buffer size is 2 KB.
Question: How to properly detect the absence of network on the service, or timeout properly, or cancel the request
The service code:
public static async Task<List<(Guid, string)>> StreamFileAsync(
this HttpRequest request, DeviceId deviceId, FileTransferInfo transferInfo)
{
var boundary = GetBoundary(MediaTypeHeaderValue.Parse(request.ContentType), DefaultFormOptions.MultipartBoundaryLengthLimit);
var reader = new MultipartReader(boundary, request.Body);
var section = await reader.ReadNextSectionAsync(_cancellationToken);
if (section != null)
{
var fileSection = section.AsFileSection();
var targetPath = transferInfo.FileTempPath;
try
{
using (var outfile = new FileStream(transferInfo.FileTempPath, FileMode.Append, FileAccess.Write, FileShare.None))
{
var buffer = new byte[DefaultCopyBufferSize];
int read;
while ((read = fileSection.FileStream.Read(buffer, 0, buffer.Length)) > 0) // HANGS HERE
{
outfile.Write(buffer, 0, read);
transferInfo.BytesSaved = read + transferInfo.BytesSaved;
}
}
}
catch (Exception e)
{
...
}
}
}
The client code:
var request = CreateRequest(fileTransferId, boundary, header, footer, filePath, offset, headers, null);
using (Stream formDataStream = request.GetRequestStream())
{
formDataStream.ReadTimeout = 60000;
formDataStream.Write(Encoding.UTF8.GetBytes(header), 0, header.Length);
byte[] buffer = new byte[2048];
using (FileStream fs = new FileStream(filePath, FileMode.Open, FileAccess.Read, FileShare.Read))
{
fs.Seek(offset, SeekOrigin.Begin);
for (int i = 0; i < fs.Length - offset;)
{
int k = await fs.ReadAsync(buffer, 0, buffer.Length);
if (k > 0)
{
await Task.Delay(100);
await formDataStream.WriteAsync(buffer, 0, k);
}
i = i + k;
}
}
formDataStream.Write(footer, 0, footer.Length);
}
var uploadingResult = request.GetResponse() as HttpWebResponse;
private static HttpWebRequest CreateRequest(
Guid fileTransferId,
string boundary,
string header,
byte[] footer,
string filePath,
long offset,
NameValueCollection headers,
Dictionary<string, string> postParameters)
{
var url = $"{_BaseAddress}v1/ResumableUpload?fileTransferId={fileTransferId}";
HttpWebRequest request = (HttpWebRequest) WebRequest.Create(url);
request.Method = "POST";
request.ContentType = "multipart/form-data; boundary=\"" + boundary + "\"";
request.UserAgent = "Agent 1.0";
request.Headers.Add(headers); // custom headers
request.Timeout = 120000;
request.KeepAlive = true;
request.AllowReadStreamBuffering = false;
request.ReadWriteTimeout = 120000;
request.AllowWriteStreamBuffering = false;
request.ContentLength = CalculateContentLength(filePath, offset, header, footer, postParameters, boundary);
return request;
}
What I tried:
I added these in into config files:
Tried to set timeout on the server
var host = new WebHostBuilder().UseKestrel(o => { o.Limits.KeepAliveTimeout = TimeSpan.FromMinutes(2);})
Used async and non-async Read()
Tried with keep alive and without
Tried to abort the request when network was restored: request?.Abort();
Tried to set formDataStream.ReadTimeout = 60000;
Since I did not find a better way, I decided to add a timeout to the reading stream and saving it to the file. The good example was posted here: https://blogs.msdn.microsoft.com/pfxteam/2012/10/05/how-do-i-cancel-non-cancelable-async-operations/
public static async Task<List<(Guid, string)>> StreamFileAsync(this HttpRequest request, DeviceId deviceId, FileTransferInfo transferInfo)
{
var boundary = GetBoundary(MediaTypeHeaderValue.Parse(request.ContentType), DefaultFormOptions.MultipartBoundaryLengthLimit);
var reader = new MultipartReader(boundary, request.Body);
var section = await reader.ReadNextSectionAsync(_cancellationToken);
if (section != null)
{
var fileSection = section.AsFileSection();
var targetPath = transferInfo.FileTempPath;
try
{
await SaveMyFile(...);
}
catch (OperationCanceledException){...}
catch (Exception){...}
}
}
private static async Task SaveMyFile(...)
{
var cts = CancellationTokenSource.CreateLinkedTokenSource(myOtherCancellationToken);
cts.CancelAfter(streamReadTimeoutInMs);
var myReadTask = StreamFile(transferInfo, fileSection, cts.Token);
await ExecuteMyTaskWithCancellation(myReadTask, cts.Token);
}
private static async Task<T> ExecuteMyTaskWithCancellation<T>(Task<T> task, CancellationToken cancellationToken)
{
var tcs = new TaskCompletionSource<bool>();
using (cancellationToken.Register(s => ((TaskCompletionSource<bool>) s).TrySetResult(true), tcs))
{
if (task != await Task.WhenAny(task, tcs.Task))
{
throw new OperationCanceledException(cancellationToken);
}
}
return await task;
}
private static async Task<bool> StreamFile(...)
{
using (var outfile = new FileStream(transferInfo.FileTempPath, FileMode.Append, FileAccess.Write, FileShare.None))
{
var buffer = new byte[DefaultCopyBufferSize];
int read;
while ((read = await fileSection.FileStream.ReadAsync(buffer, 0, buffer.Length, token)) > 0)
{
if (token.IsCancellationRequested)
{
break;
}
await outfile.WriteAsync(buffer, 0, read);
transferInfo.BytesSaved = read + transferInfo.BytesSaved;
}
return true;
}
}

Trouble writing zip content to device. Writing as hexa data not xml

I'm trying to download a zip file that contains an xml file and save the xml file to the device.
I'm able to save the file inside the zip but when I open the file it contains hexa data and not xml.
What I'm doing:
I'm using http client to make a get call. I get stream content from the response and using dependency injection, I pass the stream to a method that writes to the device.
In share code:
public static async Task<bool> Download()
{
var url = "some url";
HttpClientHandler handler = new HttpClientHandler()
{
AutomaticDecompression = DecompressionMethods.GZip | DecompressionMethods.Deflate
};
var httpClient = new HttpClient(handler);
try
{
var response = await httpClient.GetAsync(url);
var total = response.Content.Headers.ContentLength.HasValue ? response.Content.Headers.ContentLength.Value : -1L;
var stream = await response.Content.ReadAsStreamAsync();
DependencyService.Get<IFileHelper>().WriteToFilePath("download", stream, total);
}
catch (Exception e)
{
//Handle exception
}
return true;
}
in iOS:
public async void WriteToFilePath(string filename, Stream responseStream, long contentSize)
{
var downloadFilePath = GetLocalFilePath(filename);
int receivedBytes = 0;
using (FileStream fileOut = File.Create(downloadFilePath))
{
using (Stream stream = responseStream)
{
byte[] buffer = new byte[4096];
for (;;)
{
int bytesRead = await stream.ReadAsync(buffer, 0, buffer.Length);
if (bytesRead == 0)
{
await Task.Yield();
break;
}
else
{
System.Diagnostics.Debug.WriteLine(bytesRead);
fileOut.Write(buffer, 0, bytesRead);
}
receivedBytes += bytesRead;
System.Diagnostics.Debug.WriteLine("receivedBytes: " + receivedBytes + " contentSize: " + contentSize);
}
}
}

How get upload progress?

Sorry for my english. I have a method in which I send the StorageFile to the server. I tried using Windows.Web.Http.HttpClient, but does not work ( getting an invalid response from the server ) , so I use System.Net.Http.HttpClient.
Here is my code :
public static async void Upload(string uri, StorageFile data, Action<double> progressCallback = null)
{
try
{
byte[] fileBytes =await ReadFile(data);
System.Net.Http.HttpClient client = new System.Net.Http.HttpClient();
MultipartContent content = new System.Net.Http.MultipartFormDataContent();
var file1 = new ByteArrayContent(fileBytes);
file1.Headers.ContentDisposition = new ContentDispositionHeaderValue("form-data")
{
Name = "file1",
FileName = data.Name,
};
content.Add(file1);
System.Net.Http.HttpResponseMessage response = await client.PostAsync(uri, content);
response.EnsureSuccessStatusCode();
var raw_response = await response.Content.ReadAsByteArrayAsync();
var r2 = Encoding.UTF8.GetString(raw_response, 0, raw_response.Length);
if (r2[0] == '\uFEFF')
{
r2 = r2.Substring(1);
}
Logger.Info(r2);
}
catch (Exception exc)
{
Logger.Error( exc);
}
}
Whether it is possible to change the code to receive progress about downloading a file in callback function?
On Windows Runtime you can try to switch to Windows.Web.HttpClient class. Its PostAsync method returns IAsyncOperationWithProgress<HttpResponseMessage, HttpProgress> interface. This interface has Progress event to which you can simply subscribe before awaiting the result.
Simplest way to upload file with progress
I had the same issue, and after some tries found out that you can easily get byte-accurate progress by tracking the Position of the FileStream of the file that you are going to upload.
This is a sample code that shows that.
FileStream fileToUpload = File.OpenRead(#"C:\test.mp3");
HttpContent content = new StreamContent(fileToUpload);
HttpRequestMessage msg = new HttpRequestMessage{
Content=content,
RequestUri = new Uri(--yourUploadURL--)
}
bool keepTracking = true; //to start and stop the tracking thread
new Task(new Action(() => { progressTracker(fileToUpload, ref keepTracking); })).Start();
var result = httpClient.SendAsync(msg).Result;
keepTracking = false; //stops the tracking thread
And define progressTracker() as
void progressTracker(FileStream streamToTrack, ref bool keepTracking)
{
int prevPos = -1;
while (keepTracking)
{
int pos = (int)Math.Round(100 * (streamToTrack.Position / (double)streamToTrack.Length));
if (pos != prevPos)
{
Console.WriteLine(pos + "%");
}
prevPos = pos;
Thread.Sleep(100); //only update progress every 100ms
}
}
And this solved my problem.

Categories