Develop Reference

Delete file after returning it on an API request

I get a request with which I create a file and return it to the client.
After the file is sent I want it deleted.
Since I get many request, files are big and memory is scarce, I don't want to buffer it in memory to send it.
The only method I got to work without buffering the whole file in memory was:
Response.TransmitFile(filepath)
The problem with this is that it does it asynchronously, so if I delete it after that call the file download is interrupted.
I tried calling Flush, adding the delete on a finally block but neither of those worked. I thought of inheriting HttpResponse to try modify TransmitFile, but it's a sealed class. I tried to use HttpResponse.ClientDisconnectedToken but either I don't understand how to use it correctly or it isn't working in this case.
How can I achieve this? Is there a better method than calling HttpResponse's TransmitFile? Always taking into account that this is an API, files can't be broken into different requests and that it doesn't load the full file in memory.
I'm not sure if it could help somehow, but my controller is inheriting from AbpApiController.

You create the file in a temp folder, just create a job to remove all files based on date/time. Maybe give the user 4 hours to download the file.

An implementation of TransmitFileWithLimit, It can be improve in many ways, but it works
Extension for HttpResponse
public static class HttpResponseExtensions
{
public static void TransmitFileWithLimit(this HttpResponse response, string path, int limit)
{
var buffer = new byte[limit];
var offset = 0;
response.ClearContent();
response.Clear();
response.ContentType = "text/plain";
using (var fileStream = File.OpenRead(path))
{
var lengthStream = fileStream.Length;
while (offset < lengthStream)
{
var lengthBytes = fileStream.Read(buffer, 0, limit);
var chars = System.Text.Encoding.ASCII.GetString(buffer, 0, lengthBytes).ToCharArray();
response.Write(chars, 0, lengthBytes);
offset += lengthBytes;
}
}
response.Flush();
response.End();
}
}
Inside Controller
public void Get()
{
var path = #"C:\temporal\bigfile.mp4";
var response = HttpContext.Current.Response;
response.ClearContent();
response.Clear();
response.AddHeader("Content-Disposition", "inline; filename=" + HttpUtility.UrlPathEncode(path));
response.ContentType = "text/plain";
response.AddHeader("Content-Length", new FileInfo(path).Length.ToString());
response.Flush();
HttpContext.Current.ApplicationInstance.CompleteRequest();
TransmitFileWithLimit(path, 10000);
File.Delete(path);
}

What you can do is to create a byte array from your file and then delete the file.
So, once it is created, you store it in memory, then you delete the file and return the byte array.
Check this answer to see an example.
I hope this helps.

Unable to read video file greater than 2gb

In an ASP.NET application, I am using the code shown below to read video files from a shared path and play them in the browser. It is working for files which are less than 300 mb in size, but it is throwing out of memory exceptions for 650 mb file and for greater than 2 gb it is throwing
The file is too long. This operation is currently limited to supporting files less than 2 gigabytes in size.
My code:
ImpersonationHelper.Impersonate(ConfigurationManager.AppSettings["Domain"], ConfigurationManager.AppSettings["UserName"],ConfigurationManager.AppSettings["Password"], delegate
{
FileBuffer = System.IO.File.ReadAllBytes(filepath);
if (FileBuffer != null)
{
Response.Buffer = true;
Response.ContentType = contenttype;
Response.AddHeader("Content-Disposition", "inline");
Response.AddHeader("content-length", FileBuffer.Length.ToString());
Response.BinaryWrite(FileBuffer);
Response.Flush();
HttpContext.Current.ApplicationInstance.CompleteRequest();
Response.Close();
Response.End();
}
});

Loading entire file into memory is bad idea. Instead of using FileBuffer array you can do below:
// ...
using (var stream = new FileStream(filepath, FileMode.Open, FileAccess.Read))
{
Response.AddHeader("content-length", stream.Length.ToString());
stream.CopyTo(Response.OutputStream);
}
You should also remove this line Response.Buffer = true; and add Response.BufferOutput = false; in the beginning. With BufferOutput == true web server saves entire output to memory and only after this sends it to client. We don't want this to happen.

Generic handler file download does not start

I'm trying to start downloading files from server, now just with some hardcoded values for files which exists but for some reason the download does not start and no error is thrown.
This is the code I have:
public void ProcessRequest(HttpContext context)
{
string destPath = context.Server.MapPath("~/Attachments/cover.txt");
// Check to see if file exist
FileInfo fi = new FileInfo(destPath);
if (fi.Exists)
{
HttpContext.Current.Response.ClearHeaders();
HttpContext.Current.Response.ClearContent();
HttpContext.Current.Response.AppendHeader("Content-Length", fi.Length.ToString());
HttpContext.Current.Response.ContentType = "application/octet-stream";
HttpContext.Current.Response.AppendHeader("Content-Disposition", "attachment; filename=" + "cover.txt");
HttpContext.Current.Response.BinaryWrite(ReadByteArryFromFile(destPath));
HttpContext.Current.Response.End();
}
}
public bool IsReusable
{
get
{
return false;
}
}
private byte[] ReadByteArryFromFile(string destPath)
{
byte[] buff = null;
FileStream fs = new FileStream(destPath, FileMode.Open, FileAccess.Read);
BinaryReader br = new BinaryReader(fs);
long numBytes = new FileInfo(destPath).Length;
buff = br.ReadBytes((int)numBytes);
return buff;
}
I'm stepping in the code and no problem is occurring but as well no file download popup is shown in the browser.
Do you see anything wrong?

I believe the issue your having is that your calling HttpContext.Current. Since your utilizing a Generic Handler File I believe you'll want to utilize the context parameter being passed to your method signature. An example would be:
public void ProcessRequest (HttpContext context)
{
// Build Document and Zip:
BuildAndZipDocument();
// Context:
context.Response.ContentType = "application/zip";
context.Response.AddHeader("content-disposition", "filename="Commodity.zip");
zip.Save(context.Response.OutputStream);
// Close:
context.Response.End();
}
I believe if you utilize context, rather than HttpContext.Current it will resolve your issue.

NEVER use "application/zip" in your Content-Type header for a ZIP file. NEVER! Confirmed bug is in IE6 which will corrupt the download.
If you want the most universal behavior for a binary file across the most browsers past and present ALWAYS use "application/octet-stream" just like you see it used everywhere else!!
header('Content-Type: application/octet-stream');
This overcomes the IE6 bug, assuming you care. Nonetheless, you achieve nothing by switching from application/octet-stream to application/zip, so you might as well stop wasting your time on that one and keep it application/octet-stream.

RedirectToAction does not work

The RedirectToAction does not display the View.
// Go populate and display PDF using XML file
DoPDF(stXML);
}
UpDateDropDown(model);
return RedirectToAction("ReportsSelection", "Reports");
Rendering Code:
private void DoPDF(String stXML)
{
string filename = string.Concat(Guid.NewGuid().ToString(), ".pdf");
PdfReader reader = new PdfReader(new RandomAccessFileOrArray(Request.MapPath(_NFCPage._NFReference.FM_NOFEAR_PDF)), null);
// Create the iTextSharp document
// Set the document to write to memory
using (MemoryStream memStream = new MemoryStream())
{
PdfStamper ps = new PdfStamper(reader, memStream);
// Populate the PDF with values in the XML file
AcroFields af = ps.AcroFields;
ParserXML(stXML, af);
ps.FormFlattening = false;
ps.Writer.CloseStream = false;
ps.Close();
byte[] buf = new byte[memStream.Position];
memStream.Position = 0;
memStream.Read(buf, 0, buf.Length);
// Set the appropriate ContentType
Response.ContentType = "Application/pdf";
// Get the physical path to the file
Response.AddHeader("Content-disposition", string.Format("attachment; filename={0};", filename));
// Write the file directly to the HTTP content output stream.
Response.Buffer = true;
Response.Clear();
Response.BinaryWrite(memStream.GetBuffer()); //Comment out to work
Response.End(); //Comment out to work
}
}
I have noticed that if I remove the last two lines in the DoPDF routine that it does display the view.

Response.End() will cause the server to send the HTTP response. Your browser at that point will consider the request completed and the redirect won't happen. Can you provide more context on what you're trying to accomplish? Then we can get a better idea of how to help you.

Edit: nvm, you have a Response.End() call there, that will end execution of the request and your redirect will obviously not work. If you're trying to flush the stream then you need to do Response.Flush() instead.

Don't handle files downloads in MVC, as you can see, it can cause problems...
return File(memStream, "Application/pdf", filename);
Will do everything for you.
MSDN

Will HttpResponse.Filter buffer the whole data before start the sending?

An user posts this article about how to use HttpResponse.Filter to compress large amounts of data. But what will happen if I try to transfer a 4G file? will it load the whole file in memory in order to compress it? or otherwise it will compress it chunk by chunk?
I mean, I'm doing this right now:
public void GetFile(HttpResponse response)
{
String fileName = "example.iso";
response.ClearHeaders();
response.ClearContent();
response.ContentType = "application/octet-stream";
response.AppendHeader("Content-Disposition", "attachment; filename=" + fileName);
response.AppendHeader("Content-Length", new FileInfo(fileName).Length.ToString());
using (FileStream fs = new FileStream(Path.Combine(HttpContext.Current.Server.MapPath("~/App_Data"), fileName), FileMode.Open))
using (DeflateStream ds = new DeflateStream(fs,CompressionMode.Compress))
{
Byte[] buffer = new Byte[4096];
Int32 readed = 0;
while ((readed = ds.Read(buffer, 0, buffer.Length)) > 0)
{
response.OutputStream.Write(buffer, 0, readed);
response.Flush();
}
}
}
So at the same time I'm reading, I'm compressing and sending it. Then I wanna know if HttpResponse.Filter do the same thing, or otherwise it will load the whole file in memory in order to compress it.
Also, I'm a little bit insecure about this... maybe is needed to load the whole file in memory to compress it... is it?
Cheers.

HttpResponse.Filter is a stream: You can write in it in chunks.
You are doing it correctly. You are using FileStream and DeflateStream to read from file and compress it. You are reading 4096 bytes at time and then writing them into the Response Stream. So all you are using is 4096 byte(and a little bit more) of memory.