Question

I use the following code to send files from my server to the client:

Response.AppendHeader("content-disposition", "attachment; filename=" + FileName);
Response.ContentType = MimeType;
Response.WriteFile(PathToFile);
Response.End();

This works fine. Problem is, that when I download files from Internet Explorer, special characters, like the danish æ, ø and å, gets interpreted wrong. So i file with the name ‘Test æ ø å file.txt’ downloads as ‘Test æ_ø_Ã¥ file.txt’

I´ve tried adding Byte Order Mark to the response:

byte[] BOM = { 0xEF, 0xBB, 0xBF };
Response.BinaryWrite(BOM);

And setting the charset:

Response.Charset = "UTF-8";
Response.ContentEncoding = System.Text.Encoding.UTF8;

But none if it helped.
This seems only to be a problem in Internet Explorer.

Answer 1

The headers doesn’t use the encoding of the content, so it won’t help to change the content encoding. There is a standard for specifying encoding for the header parameters, but it’s pretty new, so it won’t be fully supportred in all browsers:

http://greenbytes.de/tech/webdav/rfc5987.html

This is how you would encode the name:

string FileName = "Test æ ø å file.txt";

string name = String.Concat(Encoding.UTF8.GetBytes(FileName).Select(b => {
  if ((b >= 48 && b <= 57) || (b >= 65 && b <= 90) || (b >= 97 && b <= 122)) {
    return new String((char)b, 1);
  } else {
    return String.Format("%{0:x2}", b);
  }
}).ToArray());

Response.AppendHeader("content-disposition", "attachment; filename*=UTF-8''" + name);

The text ending up in the header would look like this:

attachment; filename*=UTF-8''Test%20%c3%a6%20%c3%b8%20%c3%a5%20file%2etxt

Note: The code encodes all characters except alphanumeric. That works, but not all other characters need encoding. You could evolve the code to leave a few more characters unencoded, after checking the standards.