Fix: Properly handle node ownership when using appendChild and insertBefore

[SrikanthKumarC]

Fixes a bug where using appendChild or insertBefore with a node from a different ownerDocument would throw a WrongDocument Error. According to the specification, the new node being added should be implicitly adopted by the parent's ownerDocument.

This change updates these methods to first set the ownerDocument of the node being added, and recursively update the ownerDocument of its children to match that of the parent.

---

Currently, the behavior is as follows: an error is thrown when a node is inserted using appendChild or insertBefore if its ownerDocument differs from that of the parent. This PR fixes that.

Debug log:

js : function call error . . . . . . . . . . . . . . .  �[0m[+0ms]
      name = browser.dom.node.Node._insertBefore
      err = WrongDocument
      args = 
        1: #<HTMLDivElement> (object)
        2: #<HTMLParagraphElement> (object)
      stack = 
        <anonymous>:28

Javascript used:

console.log("--- Starting Automatic Cross-Document Insertion ---");


const parser = new DOMParser();
const newDoc = parser.parseFromString('<div id="new-node"><p>Hey</p></div>', 'text/html');
console.log("A new document has been created in memory using DOMParser:", newDoc);


const newNode = newDoc.getElementById('new-node');

console.log("The new node's initial ownerDocument is:", newNode.ownerDocument);
console.log("The main document is:", document);


const parent = document.getElementById('target-container');
const referenceNode = document.getElementById('reference-node');


parent.insertBefore(newNode, referenceNode);

console.log("\n--- Insertion Complete ---");
console.log("The new node has been inserted into the main document.");
console.log("The new node's updated ownerDocument is:", newNode.ownerDocument);
console.log("Is the new node now a child of the main document's parent?", parent.contains(newNode));

const k = document.getElementById('new-node');
const ptag = k.querySelector('p');
console.log("The new node's p tag owner is: ", ptag.ownerDocument);
console.log("P tags owner is same as main document?", ptag?.ownerDocument === document);

const statusMessage = document.getElementById('status-message');
statusMessage.textContent = "Success! The node was inserted automatically on load. Check the console for details.";

[github-actions]

All contributors have signed the CLA ✍️ ✅
_{Posted by the CLA Assistant Lite bot.}

[SrikanthKumarC]

I have read the CLA Document and I hereby sign the CLA

[krichprollsch]

Hello @SrikanthKumarC

Thank you very much for you PR 🙏

I would like to avoid the usage of an arena for traversing children.
Can you consider using the walker instead?

const w = Walker{};
while (true) {
    next = try w.get_next(child, next) orelse break;
    next.owner = self_owner;
}

existing example: https://github.com/lightpanda-io/browser/blob/main/src/browser/dom/node.zig#L224-L234

[SrikanthKumarC]

Oh yeah that's better. Just refactored it to use the walker instead. thanks

[krichprollsch]

Thanks for the change.

Here is a little non-blocking suggestion.

It would be also great to:

• add a comment before both if statements to explain the intention of changing the owner
• write a JS test case at the end of the file, similar to https://github.com/lightpanda-io/browser/blob/main/src/browser/polyfill/webcomponents.zig#L39-L65 You could use Browser.DOM.node.owner as test name. You can run it specifically with T=Browser.DOM.node.owner make test

[krichprollsch]

what about:

Suggested change

	child.owner = self_owner;
	var current = child;
	while (try w.get_next(child, current)) |current_node| {
	current_node.owner = self_owner;
	current = current_node;
	}
	var current = child;
	while (true) {
	current.owner = self_owner;
	current = try w.get_next(child, next) orelse break;
	}

[SrikanthKumarC]

Thanks for the suggestion. I’ve added the tests and simplified the walker as you recommended. Sorry for the trouble, I'm still getting the hang of Zig

[krichprollsch]

No worries, thank you very much for your contribution 🙏
I think a will create a little additional PR to avoid getting the owner twice when calling __insertBefore with a null ref_node_.