|
/// <summary>
/// IsW3CCompliant
/// http://w3c.org
/// The valid XML characters and character ranges (hex values) as defined by the W3C XML language specifications 1.0:
/// #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD] | [#x10000-#x10FFFF]
/// </summary>
/// <param name="c"></param>
/// <returns></returns>
static private bool IsW3CCompliant(char c)
{
int charInt = Convert.ToInt32(c);
return charInt == 9 || charInt == 10 || charInt == 13 || (charInt >= 32 && charInt <= 55295) || (charInt >= 57344 && charInt <= 65533) || (charInt >= 65536 && charInt <= 1114111);
}
/// <summary>
/// Scrub Xml ensures that each character is W3C compliant.
/// This is a major performance hit. . .
/// #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD] | [#x10000-#x10FFFF]
/// </summary>
/// <param name="xmlStr"></param>
/// <returns></returns>
static internal string ScrubString(string xmlStr)
{
if (xmlStr == string.Empty || xmlStr == null)
return xmlStr;
string pattern = @"[^\w\.@-]";
StringBuilder strB = new StringBuilder(xmlStr.Length);
//-- If there are no special chars just return the original (99%)
Regex regex = new Regex(pattern);
if (!regex.Match(xmlStr).Success)
return xmlStr;
char[] charArray = xmlStr.ToCharArray();
for (int i = 0; i < charArray.Length; i++)
{
if (IsW3CCompliant(charArray[i]))
{
strB.Append(charArray[i]);
}
}//for
return strB.ToString();
}
/// <summary>
/// XmlTextWriter override that scrubs the attributes
/// </summary>
public class MyXmlTextWriter : XmlTextWriter
{
/// <summary>
/// MyXmlTextWriter scrubs attribute values. Constructor just calls base.
/// </summary>
/// <param name="tw"></param>
public MyXmlTextWriter(System.IO.TextWriter tw) : base(tw) { }
/// <summary>
/// MyXmlTextWriter scrubs attribute values. Constructor just calls base.
/// </summary>
/// <param name="filename"></param>
/// <param name="encoding"></param>
public MyXmlTextWriter(string filename, System.Text.Encoding encoding) : base(filename, encoding) { }
/// <summary>
/// MyXmlTextWriter scrubs attribute values. Constructor just calls base.
/// </summary>
/// <param name="w"></param>
/// <param name="encoding"></param>
public MyXmlTextWriter(System.IO.Stream w, System.Text.Encoding encoding) : base(w, encoding) { }
/// <summary>
/// This class scrubs the attribute value - the sole
/// reason for this class.
/// </summary>
new public void WriteAttributeString(string localName, string value)
{
base.WriteAttributeString(localName, XmlHelper.ScrubString(value));
}
/// <summary>
/// This class scrubs the attribute value - the sole
/// reason for this class.
/// </summary>
public void WriteAttributeStringRaw(string localName, string value)
{
base.WriteAttributeString(localName, value);
}
}//MyXmlTextWriter
| 