What about an XPathEvaluator ?

[Siltaar]

To connect #73 and #143 I can testify that it exists real-life use-cases for XPathEvaluators, such as reaching namespaced elements in an RSS file.

The .evaluate() function is provided on DOM nodes in Firefox :
https://developer.mozilla.org/fr/docs/Web/API/Document/evaluate

That would be great to have it here also.

Going through nearly one thousand website scrapping I identified some cases where XPath look mandatory to me :

• search in ancestors of a node
• going through a flat structure (with following-sibling)
• selecting a node from its content
• specifying the 2nd text child of a node
• and even getting some content out of HTML comments

The good news is that there is nothing worst behind XPath. I don't have use-cases for something fancier or weirder… after HTML and RSS we hardly encounter worse than JSON on the web, except for multipart/form-data content-types in some POST requests.

[b-fuze]

I am not opposed to this, I'd just need to see what existing implementations are out there