JavaDoc states: "Evaluates the specified XPath expression from this node, returning the matching elements."
That implies, that with given html
<html>
<body>
<h1>Wrong</h1>
<div>
<h1>Right</h1>
</div>
<body>
</html>
and selected div node, in order to select child h1 node we need to pass xpath as //h1. But that's not the case, we need to select current node first with a dot selector, so correct xpath is .//h1. While it is proper xpath, I'd argue, that JavaDoc implies, that the node is already selected. It's specially confusing, if you print node as xml and try to validate your xpath via third party tools.
It's a bit against common sense, that selected node does not traverse from it's location. I do not expect change in the code, but more specific JavaDoc would be definitely helpful.
JavaDoc states: "Evaluates the specified XPath expression from this node, returning the matching elements."
That implies, that with given html
and selected div node, in order to select child h1 node we need to pass xpath as //h1. But that's not the case, we need to select current node first with a dot selector, so correct xpath is .//h1. While it is proper xpath, I'd argue, that JavaDoc implies, that the node is already selected. It's specially confusing, if you print node as xml and try to validate your xpath via third party tools.
It's a bit against common sense, that selected node does not traverse from it's location. I do not expect change in the code, but more specific JavaDoc would be definitely helpful.