/**
* Detects encoding from a stream.
*
* @param inp the stream
* @return a Reader with the deduced encoding.
* @throws IOException if IO went wrong
* @throws UnsupportedEncodingException if unsupported encoding was detected
*/
virtual public StreamReader DetectEncoding(Stream inp)
{
byte[] b4 = new byte[4];
int count = inp.Read(b4, 0, b4.Length);
if (count != 4)
{
throw new IOException("Insufficient length");
}
String encoding = XMLUtil.GetEncodingName(b4);
String decl = null;
int bytesNumber = 0;
if (encoding.Equals("UTF-8"))
{
StringBuilder sb = new StringBuilder();
int c;
while (bytesNumber < 1028 && ((c = inp.ReadByte()) != -1))
{
if (c == '>')
{
break;
}
sb.Append((char)c);
bytesNumber++;
}
decl = sb.ToString();
}
else if (encoding.Equals("CP037"))
{
MemoryStream bi = new MemoryStream();
int c;
while (bytesNumber < 1028 && ((c = inp.ReadByte()) != -1))
{
if (c == 0x6e) // that's '>' in ebcdic
{
break;
}
bi.WriteByte((byte)c);
bytesNumber++;
}
decl = Encoding.GetEncoding(37).GetString(bi.ToArray());
}
if (decl != null)
{
decl = EncodingUtil.GetDeclaredEncoding(decl);
if (decl != null)
{
encoding = decl;
}
}
if (inp.CanSeek)
{
inp.Seek(0, SeekOrigin.Begin);
}
return(new StreamReader(inp, IanaEncodings.GetEncodingEncoding(encoding)));
}
/**
* Parses the XML document firing the events to the handler.
* @param doc the document handler
* @param in the document. The encoding is deduced from the stream. The stream is not closed
* @throws IOException on error
*/
public static void Parse(ISimpleXMLDocHandler doc, Stream inp)
{
byte[] b4 = new byte[4];
int count = inp.Read(b4, 0, b4.Length);
if (count != 4)
{
throw new IOException(MessageLocalization.GetComposedMessage("insufficient.length"));
}
String encoding = XMLUtil.GetEncodingName(b4);
String decl = null;
if (encoding.Equals("UTF-8"))
{
StringBuilder sb = new StringBuilder();
int c;
while ((c = inp.ReadByte()) != -1)
{
if (c == '>')
{
break;
}
sb.Append((char)c);
}
decl = sb.ToString();
}
else if (encoding.Equals("CP037"))
{
MemoryStream bi = new MemoryStream();
int c;
while ((c = inp.ReadByte()) != -1)
{
if (c == 0x6e) // that's '>' in ebcdic
{
break;
}
bi.WriteByte((byte)c);
}
decl = Encoding.GetEncoding(37).GetString(bi.ToArray());//cp037 ebcdic
}
if (decl != null)
{
decl = GetDeclaredEncoding(decl);
if (decl != null)
{
encoding = decl;
}
}
Parse(doc, new StreamReader(inp, IanaEncodings.GetEncodingEncoding(encoding)));
}