page: http://h4z.it/View/-20100729_designtea.jpg
code to execute in console:
document.evaluate("//img",document,null,9,null).singleNodeValue
or
document.evaluate("//a",document,null,9,null).singleNodeValue
or even
document.evaluate("//html",document,null,9,null).singleNodeValue
result (tested both with Chrome and Firefox): null
I thought that page overrided document.evaluate but it shows
document.evaluate
function evaluate() { [native code] }
and delete document.evaluate does not help, so what else can it be which breaks document.evaluate ?
The page you show in your question uses an xhtml namespace and probably so do the other pages that you are seeing this happen on.
Since you are setting null for the namespace resolver argument it cannot find the elements.
Note: XPath defines QNames without a prefix to match only elements in the null namespace. There is no way in XPath to pick up the default namespace as applied to a regular element reference (e.g., p[@id='_myid'] for xmlns='http://www.w3.org/1999/xhtml'). To match default elements in a non-null namespace, you either have to refer to a particular element using a form such as ['namespace-uri()='http://www.w3.org/1999/xhtml' and name()='p' and @id='_myid'] (this approach works well for dynamic XPath's where the namespaces might not be known) or use prefixed name tests, and create a namespace resolver mapping the prefix to the namespace.
So if you setup a resolver than you can access the element correctly by prefixing them:
function resolver(prefix){
return prefix === "xhtml" ? 'http://www.w3.org/1999/xhtml' : null;
}
document.evaluate("//xhtml:a",document,resolver,9,null).singleNodeValue
If you want to get a node without having to know the namespace you can use local-name() XPath function as part of the expression
document.evaluate("//*[local-name()='img']",document,null,9,null).singleNodeValue