Marin Atanasov

> Regarding `.firstChild` and testing in general, I think the way we're testing this component is philosophically completely wrong. The component doesn't have any UI. It's a form with fields...

Thanks for your feedback everyone! > I'm unable to explain why, but I am seeing small flashes of content behind the loader: Yes, those are easy to reproduce with CPU...

> We might want to identify some structural parts like editor header and navigation bits (W menu) first and display those as soon as ready so you can go elsewhere...

Thanks for the demo! At a first glance, this would be a bit troublesome with the current interface skeleton structure. Let me clarify. Currently, we have this rough structure for...

Thanks for the feedback everyone! I've done some changes and here's how it looks now: https://user-images.githubusercontent.com/8436925/185119864-4cf1474e-2816-4073-8b9c-2c0d3cc6c54c.mov Basically, the header is now outside and since it generally loads quickly and doesn't...

I plan to focus on addressing the feedback and getting it to a more shippable state next week.

@mtias @ockham I'm prioritizing this till the end of this week.

I've spent some time with this today, rebasing it and working on the feedback here. Here's a preview: https://user-images.githubusercontent.com/8436925/190165333-5b393c44-a819-48e5-a80d-9e45774102e4.mov Since the last version, now we've addressed all remaining `useSelect()` instances...

> One major downside is that all blocks, even the 3rd party ones that we don't control, need to use Suspense consistently. Otherwise, when classic and Suspense data loading is...

This one fixes the 404 link: https://github.com/christianbach/tablesorter/pull/66 And concerning the 404's of the examples - it appears that the examples on the site are not up to date with the...