Thanks Gary,
I saw your comments. You are right. So I'm looking for a solution for
this question -
I set transformation result as a SAXResult and use a customized SAX
parser to parse the result. To achieve a better performance, I want the
SAX parser to skip parsing some elements and treat them as a cdata
section. What I did is using <xsl:value-of
disable-output-escaping="yes" select="<![CDATA["/> and <xsl:value-of
disable-output-escaping="yes" select="]]>"/> to encapsulate that
section. Then I expect the SAXParser will find the <![CDATA[ and not
parse the following elements until met the closing ]]>. However, it
seems this is not achievable. Only after I serialize the output, I can
let a SAX parser to treat that section as a cdata. Am I correct here?
Any ideas?
Guoliang
--- a sample of xsl -------------------
<xsl:value-of disable-output-escaping="yes" select="<![CDATA["/>
<xsl:copy-of select="a"/>
<xsl:value-of disable-output-escaping="yes" select="]]>"/>
--- expected output -------------------
<![CDATA[<a>....</a>]]>
<a>...</a> will be treated as a text node and not be parsed by SAX parser.
----------------------------------------
> I have updated bugzilla. Basically, Xalan is working as designed. See
> the comments in bugzilla and the javadoc for the Result interface.
>
> Gary
>
> Guoliang Cao wrote:
>
>> I reported it two months ago. It's still in NEW status. If it can't be
>> fixed, I have to change my code to avoid using SAXResult. I really
>> don't like to do that.
>>
>> Sincerely,
>>
>> Guoliang
>