I have a solution that works for now, but is not using TargetCharset
from flat file assembler. I have done a custom pipeline component that
calls flat file assembler and then converts the result into
iso-8859-1:
protected override IBaseMessage Assemble(IPipelineContext pc)
{
IBaseMessage retMsg = _assembler.Assemble(pc);
System.Text.Encoding encoder =
System.Text.Encoding.GetEncoding("iso-8859-1");
byte[] buffer = new byte[1024];
byte[] destBuffer = new byte [1024];
System.IO.Stream strm = retMsg.BodyPart.Data;
VirtualStream vStrm = new VirtualStream();
while (strm.Read(buffer, 0, 1024) > 0)
{
destBuffer = System.Text.Encoding.Convert
(Encoding.UTF8, encoder, buffer);
vStrm.Write(destBuffer, 0, destBuffer.Length);
}
retMsg.BodyPart.Data = vStrm;
retMsg.BodyPart.Data.Position = 0;
return retMsg;
}
This works, I mean, after flat file assembler the output file is
correct in utf-8 (for example, N tilde is represented by bytes C3 91)
and then I can convert it to iso-8859-1 (N tilde is converted from C3
91 to D1), but if I use directly XMLNORM.TargetCharset = "iso-8859-1",
after assembler I get spanish characters as question marks (byte 3F).
I think the problem is with assembler encoding, because without set
any charset (utf-8 by default) it works correctly.