Node.childNodes return type appears to be slightly inaccurate

[dsifford]

Greetings!

Ran into this earlier today and I'm wondering if it's a bug or if I'm just mistaken for what I assumed would be returned.

Below is a snapshot of the definition of Node.

https://github.com/Microsoft/TypeScript/blob/c7b4ed3a91964915b953b34ad2ae72f36e7d6efe/lib/lib.dom.d.ts#L8260-L8265

Currently, Node.childNodes returns a NodeList which is mostly correct. However, wouldn't it be more accurate to instead return a NodeListOf<Node & ChildNode> so that the remove method is exposed?

Example of the issue:

declare const node: Node;

for (const child of node.childNodes) {
    child.remove(); // fails "Property 'remove' does not exist on type 'Node'
}

Happy to submit a PR for this if needed (I'm going to also submit one for the issue I opened yesterday when I get some spare time this weekend).

Thanks in advance 😄

[mhegazy]

@kitsonk thoughts?

[kitsonk]

The living standard says that:

The childNodes attribute’s getter must return a NodeList rooted at the context object matching only children.

and that a child is:

An object that participates in a tree has a parent, which is either null or an object, and has children, which is an ordered set of objects. An object A whose parent is object B is a child of B.

But of course there is a challenge in modelling that in Web IDL, because the types in there don't have the concept of generics, so both the IDL from Edge and the WHATWG IDL don't say anything but NodeList. So it seems totally appropriate for it to be NodeListOf<Node & ChildNode> as that is clearly more spec aligned.