Parsing errors with CDATA tags using oracle xml parser v2

Question

I'm using the oracle.xml.parser.v2 parser to combine a
generated xml document from a database with a static
xsl file to produce html. Within the xml document, there
are fairly large CDATA sections (500 lines) which happen
to contain javaScript.
Occasionally, I'm getting xml tags in the final html
document! It seems that the oracle parser pukes
on a certain type of data within a CDATA tag, and then
replaces an angle bracket (<) of a tag with a #60. This
can cause html tags being viewed in the document. At
first, I thought the amount of JavaScript within the
CDATA tag was too large for the parser to handle,
because when I split it up into 2 or more parts the errors
went away. But, re-arranging the JavaScript (String in the CDATA tag)
can make the errors disappear. And, if I
use the Saxon parser to manually combine the xsl with
the xml, the output is fine. If anyone knows that this is
a confirmed bug by Oracle, or any other relevant info on
this, please let me know.

Answer 1

Hi,
Your lucky, i've just finished a project that used the oracle parser extensively.
I think the problem may stem from your xsl. Although your cdata tags may be well formed, in the xsl you might need to escape the text again, this may mean that the xsl needs to print a further cdata tag around the data. This is because the parser (well the version I was using) strips the cdata tags before the transformation happens.
This is the probably the cause of the nasty html output.