Re: Question for implementors: XPath model and CDATA sections

The XPath specification is very clear that CDATA sections
don't exist in the XPath data model
(http://www.w3.org/TR/xpath#section-Text-Nodes).

The Xalan implementation is clearly wrong and has to be fixed.

Regards,  Martin.

At 10:14 02/02/12 +0100, Christian Geuer-Pollmann wrote:
>Hi all,
>
>while playing arounf with Xalan/Xerces, I found a - hm problem? - with 
>XPath. I read in a document which contains something like this:
>
><element>
>previous sibling text
><![CDATA[
>    text in CDATA
>]]>
>following sibling text
></element>
>
>
>and use Xalan to select all Nodes by using the c14n-with-comments expr: 
>(//. | //@* | //namespace::*). In the resulting NodeList, there is only 
>one text node (the previous sibling). The CDATA section and the following 
>sibling do not show up in the result. In the bug-report I opened for that 
>[1], there is stated by the Xalan developers:
>
>  "So when you run an XPath against a DOM and
>   return text as a DOM node, we return the
>   _first_ such node in the contiguous text,
>   whether it's Text or CDATASection. It is
>   the caller's responsibility to check whether
>   additional text follows, if that's relevant
>   for your application."
>
>So actually, I don't know whether this is right. Any opinions? I guess 
>this is a problem that can have many of the Java based implementations?
>
>Regards,
>Christian
>
>
>
>[1] http://nagoya.apache.org/bugzilla/show_bug.cgi?id=6329
>

Received on Tuesday, 12 February 2002 07:03:40 UTC