AJAX - The Complete Reference

162 Part I: Core Ideas

You then loop over each item and extract the text contents of each tag to be printed out

in a table:

for (var i=0;i<items.snapshotLength;i++)

{

var title = document.evaluate('title', items.snapshotItem(i), null,

XPathResult.FIRST_ORDERED_NODE_TYPE, null).singleNodeValue.textContent;

var url = document.evaluate('url', items.snapshotItem(i), null,

XPathResult.FIRST_ORDERED_NODE_TYPE, null).singleNodeValue.textContent;

...snip ...

/* print out the complete bookmark */

}

We omit the messy details of the output, but you can see the complete Firefox-specific

version at http://ajaxref.com/ch4/xmlxpathff.html.

Internet Explorer is different but is actually a bit cleaner looking syntax wise, though the

support is not native to the browser. Because of this, you must create an ActiveX-based

DOM parser and load the XHR’s responseText containing the bookmark list directly.

var xmlDoc = new ActiveXObject("Microsoft.XMLDOM");

/* we also could use our createDocument() wrapper function here but for

clearly acknowledging the ActiveX parsing involved we leave this here.

The online version does however use the wrapper */

xmlDoc.async="false";

xmlDoc.loadXML(xhr.responseText);

With the document parsed, you can call the selectNodes() method and pass it the

XPath expression. With the returned list of <bookmark> tags, you can then loop over the list

and call selectNodes() again and fetch the text values very directly:

FIGURE 4-7 Bookmark list populated by Ajax and XPath