Re: [dom] setAttributeNode should not change order of existing attributes (#116)

The issue is whether or not `setAttributeNode()` and its various aliases need to replace an attribute in place (behavior of Chromium/Edge/WebKit or can remove the existing attribute and then append a new attribute (behavior of Gecko/spec). The former would require a new primitive for a method that's extremely obscure, the latter doesn't.

Aside from that there's a question that if we do go with the former, what the mutation observer behavior should be. And it sounds like it should be the same as that of changing an attribute's value.

---
Reply to this email directly or view it on GitHub:
https://github.com/whatwg/dom/issues/116#issuecomment-159303185