| 06:53 | <The Freelancer> | zcorpan: I still don't understand the intuition behind this decision. I have read the resources that you pointed out, but still things are vague. There must be some reason to use NodeList for querySelectorAll(), or otherwise it was just a bad design decision... |
| 07:12 | <annevk> | I'm going to let testsuite.org expire. If anyone here wants it let me know. |
| 08:00 | <freddy> | It's a good domain name. |
| 11:39 | <Noam Rosenthal> | annevk: if you're around can you review https://github.com/whatwg/xhr/pull/347 when you have time? thanks! |
| 18:23 | <zcorpan> | The Freelancer: I don't really know, don't recall if HTMLCollection was ever proposed for querySelectorAll. Maybe at the time, HTMLCollection was seen as a HTML-specific thing while qSA can return any element, so NodeList seemed more appropriate |
| 18:24 | <zcorpan> | The Freelancer: if we were to design it today, the return type would probably be sequence<Element> |
| 18:38 | <The Freelancer> | No problem 🙂. Just thought maybe you knew of it. I don't get why browsers shifted from using NodeList for getElementsByTagName() and getElementsByClassName() to using HTMLCollection. I went through the resource that you provided from 2012. It contained a couple of chat messages related to this, but I couldn't understand its purpose still. |
| 18:39 | <The Freelancer> | zcorpan: I learnt that one thing that irritated browser developers was the fact that HTMLCollection had a namedItem() method. Do you have any knowledge about this? |