Re: document.evaluate

Hi,

Thanks for the response. I found a work around by loading the XML island 
into an XMLDocument using the DOMParser as follows:

Element.prototype.selectSingleNode = function(cXPathString, xNode)

{

//create a DOMParser

var objDOMParser = new DOMParser();

if( !xNode ) { xNode = this; }

if(xNode.innerHTML){

var oXML = objDOMParser.parseFromString(xNode.innerHTML, "text/xml");

}

else{

var oXML = objDOMParser.parseFromString(xNode.xml, "text/xml");

}

dump(oXML.documentElement.nodeName == "parsererror" ? "error while parsing" 
: oXML.documentElement.nodeName);

var xItems = this.selectNodes(cXPathString, oXML);

if(xItems.length > 0){

return xItems[0];

}

else{

return null;

}

}


"Jonas Sicking" <[email protected]> wrote in message 
news:[email protected]...
> Leon Bezuidenhout wrote:
>> I have a page with a couple of XML islands. I call the following code to 
>> return a specific node in one of the XML islands
>>     document.evaluate(cXPathString, xNode, oNSResolver, 
>> XPathResult.FIRST_ORDERED_NODE_TYPE, null);
>> where xNode is a reference to the XML island I want to query. However, my 
>> result is always in context of the first XML island on the form
>>
>> Is this a bug or am I doing something wrong?
>
> The issue is probably with the "XML island" rather than the XPath query. 
> Firefox does not support XML islands in HTML pages, the whole page will be 
> parsed using HTML parsing rules which may not produce the result you want. 
> Use the DOM-Inspector to verify that the DOM looks like what you expect.
>
> Alternatively, the problem could be in your XPath query. Note that the 
> contents in your "XML island" is part of the normal page, so if you 
> evaluate an expression like "//foo" you'll get all foo elements in the 
> entire document, not just in that "XML island". And since you're 
> requesting just the first node that node would likely be in the first 
> island.
>
> Hard to know what else could be the problem without an actual testcase.
>
> / Jonas