What character encoding is used by StreamReader.ReadToEnd()?

Question

• What character encoding is used by StreamReader.ReadToEnd()?
• What would be the reason to use (b) instead of (a) below?
• Is there a risk of their being a character encoding problem if (a) is used instead of (b)?
• Is there another method that is better than (a) and (b)?

(a)

Dim strWebResponse As String
Dim Request As HttpWebRequest = WebRequest.Create(Url)
Using Response As WebResponse = smsRequest.GetResponse()
    Using reader As StreamReader = New StreamReader(Response.GetResponseStream())
        strWebResponse = reader.ReadToEnd()
    End Using
End Using

(b)

Dim encoding As New UTF8Encoding()
Dim strWebResponse As String
Dim Request As HttpWebRequest = WebRequest.Create(Url)
Using Response As WebResponse = Request.GetResponse()
    Dim responseBuffer(Response.ContentLength - 1) As Byte
    Response.GetResponseStream().Read(responseBuffer, 0, Response.ContentLength - 1)
    strWebResponse = encoding.GetString(responseBuffer)
End Using

Solution

The standard encoding used by StreamReader is Encoding.Default, which will vary from machine to machine depending on your version of Windows and the locale that you have set. Encoding.UTF8.

I have trouble remembering what the defaults are, so I prefer to use the StreamReader constructor that lets me specify the encoding. For example:

Using reader As StreamReader = New StreamReader(Response.GetResponseStream(), Encoding.UTF8)

See the constructor documentation for more info.

If you use that constructor in your example a, the results will be the same as for your example b.

Should you use UTF-8? That depends on the page you're downloading. If the page you're downloading was encoded with UTF-8 then, yes, you should use UTF-8. UTF-8 is supposed to be the default if no character set is defined in the HTTP headers. But you need to check the Content-Type header to determine if the page uses some other encoding. For example, the Content-Type header might read:

 application/xml; charset=ISO-8859-2

You would have to examine the ContentType property of the HttpWebResponse, check to see if there is a charset field, and set the encoding properly based on that.

Or, just use UTF-8 and hope for the best.