strip invalid xml characters

今天有同事遇到了XML中包含特殊字符""，导致XML解析出错，他的IE7解析错误，我的FF3也解析出错，但我的IE6却可以显示正常，只是状态栏提示警告信息。
于是我在网上查找相关资料，发现W3C中指定不能包括这些特殊字符。

对于XML，我们一般只对以下字符进行转义(避免escape这些字符)：
"<"      "&lt;" 
">"      "&gt;"
"\""     "&quot;" 
"\'"     "&apos;" 
"&"      "&amp;"
其实这些这符，在节点文本中使用<![CDATE[]]>处理，是允许的。

Assuming your ASP is not trying to add any non-printable characters ot the XML, it usually suffices to filter and replace characters as follows:
    For any text node child of an element:
      "<"  becomes  "&lt;"
      "&"  becomes  "&amp;"

    For any attribute value:
      "<"  becomes  "&lt;"
      "&"  becomes  "&amp;"
      '"'  becomes  '&quot;' (if you are using quote(") to delimit the attribute value)
      "'"  becomes  "&apos;" (if you are using apostrophe(') to delimit the attribute value)

但是在W3C标准中只能限制以下字符才可以正确使用
http://www.w3.org/TR/2004/REC-xml-20040204/#charsets
或http://www.ivoa.net/forum/apps-samp/0808/0197.htm

XML processors MUST accept any character in the range specified for Char.

Character Range
Char    ::=    #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD] | [#x10000-#x10FFFF] /* any Unicode character, excluding the surrogate blocks, FFFE, and FFFF. */

像以下16进制字符是不允许出现在XML中的，即使放在<![CDATE[]]> 中，也不能幸免遇难。
\\x00-\\x08
\\x0b-\\x0c
\\x0e-\\x1f

按Character Range说明，除了以上3段需要排除外，另外还有一些也不能在XML中使用，像#xD800-#xDFFF，由于本人不知道这些字符是个什么样，一般应用也很难会出现这些字符，所以暂不作排除，如有需要可自行加上排除处理

简单处理c# code:
string content = "slei20sk<O?`";
content = Regex.Replace(content, "[\\x00-\\x08\\x0b-\\x0c\\x0e-\\x1f]", "*");
Response.Write(content);

网上实例代码(区别XML1.0和XML1.1，特别注意XML1.0和XML1.1不同)：
http://balajiramesh.wordpress.com/2008/05/30/strip-illegal-xml-characters-based-on-w3c-standard/
W3C has defined a set of illegal characters for use in XML . You can find info about the same here:
XML 1.0(http://www.w3.org/TR/2006/REC-xml-20060816/#charsets) | XML 1.1(http://www.w3.org/TR/xml11/#charsets)

Here is a function to remove these characters from a specified XML file:

using System;
using System.IO;
using System.Text;
using System.Text.RegularExpressions;

namespace XMLUtils
{
    class Standards
    {
        /// <summary>
        /// Strips non-printable ascii characters
        /// Refer to http://www.w3.org/TR/xml11/#charsets for XML 1.1
        /// Refer to http://www.w3.org/TR/2006/REC-xml-20060816/#charsets for XML 1.0
        /// </summary>
        /// <param name="filePath">Full path to the File</param>
        /// <param name="XMLVersion">XML Specification to use. Can be 1.0 or 1.1</param>
        private void StripIllegalXMLChars(string filePath, string XMLVersion)
        {
            //Remove illegal character sequences
            string tmpContents = File.ReadAllText(filePath, Encoding.UTF8);

            string pattern = String.Empty;
            switch (XMLVersion)
            {
                case "1.0":
                    pattern = @"#x((10?|[2-F])FFF[EF]|FDD[0-9A-F]|7F|8[0-46-9A-F]9[0-9A-F])";
                    break;
                case "1.1":
                    pattern = @"#x((10?|[2-F])FFF[EF]|FDD[0-9A-F]|[19][0-9A-F]|7F|8[0-46-9A-F]|0?[1-8BCEF])";
                    break;
                default:
                    throw new Exception("Error: Invalid XML Version!");
            }

            Regex regex = new Regex(pattern, RegexOptions.IgnoreCase);
            if (regex.IsMatch(tmpContents))
            {
                tmpContents = regex.Replace(tmpContents, String.Empty);
                File.WriteAllText(filePath, tmpContents, Encoding.UTF8);
            }
            tmpContents = string.Empty;
        }
    }
}

补上msdn上类似的处理：
http://msdn.microsoft.com/en-us/library/k1y7hyy9(vs.71).aspx
internal void CheckUnicodeString(String value)
    {
    for (int i=0; i < value.Length; ++i) {
        if (value[i] > 0xFFFD)
        {
            throw new Exception("Invalid Unicode");
        }
        else if (value[i] < 0x20 && value[i] != '\t' & value[i] != '\n' & value[i] != '\r')
        {
            throw new Exception("Invalid Xml Characters");
        }
    } 

 

 附ascii表:
http://www.asciitable.com
http://code.cside.com/3rdpage/us/unicode/converter.html