How to get ASCII encoded XML document while writing arbitrary unicode data (e.g. when transforming with XslTransform class)

One could ask - btw, can ASCII encoded XML document to contain arbitrary Unicode characters? While it sounds a bit contradictory, the answer is of course - sure it can, because encoding of an XML document (that one which you can see in XML declaration) is sort of "transfer encoding", while character range for any XML document is always the whole set of legal characters of Unicode and ISO/IEC 10646 (more strict definition). XML syntax allows any (but legal in XML) character to be written as numeric character reference, e.g. ב (Hebrew letter BET). So the solution for the given problem of narrowing the encoding is to encode all characters that don't fit into target encoding as numeric character references. Basically that's what XSLT 1.0 spec requires from XSLT processors:

It is possible that the result tree will contain a character that cannot be represented in the encoding that the XSLT processor is using for output. In this case, if the character occurs in a context where XML recognizes character references (i.e. in the value of an attribute node or text node), then the character should be output as a character reference; otherwise (for example if the character occurs in the name of an element) the XSLT processor should signal an error.

Well, unfortunately native .NET 1.X XSLT processor - XslTransform class doesn't support that yet. So let's see how we can get this done. The first solution that comes in my mind is simple custom XmlWriter, which filters output text and encodes all non ASCII characters as numeric character references. Just like SAX filter for those SAX minded. Here is the implementation:

public sealed class ASCIIXmlTextWriter : XmlTextWriter 
{
  //Constructors - add more as needed
  public ASCIIXmlTextWriter(string url) :
    base(url, Encoding.ASCII) {}

  public override void WriteString(string text)
  {
    StringBuilder sb = new StringBuilder(text.Length);
    foreach (char c in text)
    {
      if (c > 0x0080)
      {
        sb.Append("&#");
        sb.Append((int)c);
        sb.Append(';');
      }
      else
      {
        sb.Append(c);
      }
    }
    base.WriteRaw(sb.ToString());
  }
}

<?xml version="1.0" encoding="UTF-8"?>
<message>English, Русский (Russian), עברית (Hebrew)</message>

<xsl:stylesheet version="1.0" 
xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
  <xsl:template match="/">
    <xsl:copy-of select="message"/>
  </xsl:template>
</xsl:stylesheet>

XPathDocument doc = new XPathDocument("foo.xml");
XslTransform xslt = new XslTransform();
xslt.Load("foo.xsl");
ASCIIXmlTextWriter writer = new ASCIIXmlTextWriter("out.xml");
xslt.Transform(doc, null, writer, null);
writer.Close();

<message>English, &#1056;&#1091;&#1089;&#1089;&#1082;&#1080;&#1081; 
(Russian), &#1506;&#1489;&#1512;&#1497;&#1514; (Hebrew)</message>

Above simple solution handles equally well both text and attribute values, but not comments and PIs (but that's easy as well). It's neither optimized nor tested, but you've got the idea.