I have written a client to a web service. I use ASP.NET 1.1 for the
client.
It worked until I tried to send accented characters. Then the service
answered that my signature is not valid. I was told by the people
running the service that their host takes ISO-8859-1 (Latin1) and
nothing else.

I had been sending the default UTF-8. I tried to send Latin1 with this:

RequestEncoding= System.Text.Encoding.GetEncoding( "ISO-8859-1");

in the proxy class.

I tried and it seemed to work, because it had deleted all the accented
characters! Thus the service had no problem to validate the signature.
It had written <?xml version="1.0" encoding="iso-8859-1"?> and the
ContentType was the desired "text/xml; charset=iso-8859-1".

Reading from "GotDotNet User Sample: Encoding SOAP extension" (
http://www.gotdotnet.com/Community/UserSamples/Details.aspx?SampleGuid=e5ac0293-
1481-47b0-916d-5186106eb57c
) (including the little comment at the bottom), I added a
SoapExtension that goes before the rest of extensions (those that log a
copy, add a signature,...)

I did things like this:

if( SoapMessageStage.AfterSerialize== message.Stage)
{
       if( inwardStream.CanSeek)
       {
         inwardStream.Position= 0;
       }

    Regex expReg= new Regex(
        System.Text.Encoding.UTF8.HeaderName /* utf-8 */,
        RegexOptions.IgnoreCase);
    TextReader lector= new StreamReader( inwardStream);
       TextWriter escritor;
       escritor= new StreamWriter(
               // _encoding is "iso-8859-1"
        outwardStream, System.Text.Encoding.GetEncoding( _encoding));
    SoapMsgTxt= expReg.Replace( SoapMsgTxt, _encoding);
       escritor.Flush();
}

It substituted in <?xml ?> but when the next SoapExtension converted
its stream to a string, the accented characters had disappeared.

Then I tried:

escritor= new XmlTextWriter( outwardStream,
    System.Text.Encoding.GetEncoding( _encoding));
XmlDocument doc= new XmlDocument();
//SoapMsgTxt is the SOAP XML in a string, read from the stream
doc.LoadXml( SoapMsgTxt);
doc.WriteTo( escritor);

And it appeared mutilated again.

I also tried:

escritor.WriteRaw( SoapMsgTxt);

and even:

if( SoapMessageStage.BeforeSerialize== message.Stage)
{
    message.ContentType= "text/xml; charset=iso-8859-1";
       // I can call the base. I am inheriting from a subclass
       // of SoapExtension
    base.ProcessMessage( message);
}

It kept mutilating characters.

Any suggestion on how I can convert my Request to ISO-8859-1?

Thanks in advance.

--
David Mediavilla    <6kjfsyg02@sneakemail.com>