Unescaped CDATA?

Hi,

I use a Python script to build XML files out of database records. Within these
records, a certain field contains (not necessarily well-formed) HTML.

I need my XSLT processor  to skip this data when transforming the document to
an HTML document. So I thought it would be a great idea to enclose the data in
a CDATA section:

<content>
    <![CDATA[
    (malformed HTML here)
    ]]>
</content>

This works - but the tags within the CDATA section are escaped to literal <p>,
<i> etc. So the browser displays

<i>Serapions-Br|der</i>

instead of the italic text Serapions-Br|der. How can I

a) keep the XSLT processor from choking at the malformed HTML and
b) leave the tags within the HTML unescaped?

Thanks,

Jan
--
If all else fails read the instructions. - Donald Knuth