In addition to overriding React's public API, Preact Signals works by taking React internals and patching them while relying on fragile assumptions about how they work. This library is an entirely unsupported way to use React. The only reason it isn't constantly breaking is that we haven't done stable releases recently.

If React was a piece of hardware, this is exactly the kind of thing that voids the warranty. We can't help with debugging issues here, and we don't consider them bugs because so much has been meddled with.

If you have a repro case that doesn't rely on Preact Signals, we'd like to have a look. Thanks.

At least they got State Mangement right. Instead of talking about how they meddled with the ergonomics of React, isn't it better to learn from what they have done, to find a way to officially support Signals.

Its just more linear and feels natural to work with. I almost wonder why the core team had to make initial state management a pain with things like the reducer pattern which introduces a whole bunch of of issues.

If react where a design canvas on a scale of 10 it would be ranked very poorly. Ranking react at (2) would be very modest.

We all look at Signals and think useState, useReducer, useSelector, Actions, Dispatchs etc... are state management primitives the react core team designed to sound smart.

They are inefficient, long, hard, unnatural and just outright unpleasant to work with a concept like Signals completely render all those concepts redundant.

Lastly the preact team may have patched a few bits and pieces to make Signals work for the most part but at least they made State Management digestible.

And I suggest React does the same.

TLDR: The only reason we use JS is because of React and the only reason we use React is because of meta frameworks like NEXTJS
& REMIX.

I speak for most or all react developers.
Cheers!