C# - From NetworkStream to XElement managing different encoding

Question

I've an application that search XML over the network (using TcpClient), these XML have various encoding (one site in UTF8, other in Windows-1252). I would like encode all of these XML in UTF-8 (always) to be sure I'm clean.

How can I do the conversion from the NetworkStream to an XElement encoding correctly all data?

I've this :

NetworkStream _clientStream = /* ... */;
MemoryStream _responseBytes = new MemoryStream();

// serverEncoding -> Xml Encoding I get from server
// _UTF8Encoder -> Local encoder (always UTF8)

try
{
    _clientStream.CopyTo(_responseBytes);

    if (serverEncoding != _UTF8Encoder)
    {
        MemoryStream encodedStream = new MemoryStream();
        string line = null;
        using (StreamReader reader = new StreamReader(_responseBytes))
        {
            using (StreamWriter writer = new StreamWriter(encodedStream))
            {
                while ((line = reader.ReadLine()) != null)
                {
                    writer.WriteLine(
                        Encoding.Convert(serverEncoding, _UTF8Encoder, serverEncoding.GetBytes(line))
                        );
                }
            }
        }
        _responseBytes = encodedStream;
    }

    _responseBytes.Position = 0;
    using (XmlReader reader = XmlReader.Create(_responseBytes))
    {
        xmlResult = XElement.Load(reader, LoadOptions.PreserveWhitespace);
    }
}
catch (Exception ex)
{ }

Have you a better solution (and by ignoring all '\0' ?).

Edit

This works :

byte[] b = _clientStream.ReadToEnd();
var text = _UTF8Encoder.GetString(b, 0, b.Length);
xmlResult = XElement.Parse(text, LoadOptions.PreserveWhitespace);

But this not :

using (var reader = new StreamReader(_clientStream, false))
    xmlResult = XElement.Load(reader, LoadOptions.PreserveWhitespace);

I don't understand why ...

I removed code from catch clause, not interesting here, it's ok ;) — Arnaud F., Mar 22 '11 at 15:55
What happens if you create a `StreamReader` and call `ReadToEnd()`? — SLaks, Mar 23 '11 at 00:03

score 2 · Accepted Answer · answered Mar 22 '11 at 15:13

2

You can simply create a StreamReader around the NetworkStream, passing the encoding of the stream, then pass it to XElement.Load:

XElement elem
using(var reader = new StreamReader = new StreamReader(_clientStream, serverEncoding))
    elem = XElement.Load(reader);

There is no point in manually transcoding it to a different encoding.

answered Mar 22 '11 at 15:13

SLaks

868,454
176
1,908
1,964

It doesn't work. Program block on "Load" because it doesn't perform a Read operation ... But when I do : reader.BaseStream.Read() ... I got a result back... – Arnaud F. Mar 22 '11 at 15:47
That shouldn't happen. Are you sure you have the correct encoding? – SLaks Mar 22 '11 at 15:50
Yes, sure... Is the encoding blocking when not correct !? If yes, why not just throw an exception? ... – Arnaud F. Mar 22 '11 at 15:53
If you pass an encoding with multi-byte chars, it can block forever while waiting for the next byte in a multi-byte char if it's at the end of the stream. – SLaks Mar 22 '11 at 15:56
I removed encoding from StreamReader, same problem ... `using(var reader = new StreamReader(_clientStream, false)) elem = XElement.Load(reader, LoadOptions.PreserveWhitespace);` – Arnaud F. Mar 22 '11 at 16:00

C# - From NetworkStream to XElement managing different encoding

Edit

1 Answers1