jotai VS Immer

Daishi Kato, creator of Zustand and Jotai, says there are a few things we need to support RSC serialization:

I've just published jotai-wrapper, a super tiny and simple utility library that makes using jotai even simpler. It comes from the necessity to migrate a project with react-context-slices to jotai. Both libraries share a similar API. While in react-context-slices you have the useSlice hook, in jotai you have the useAtom, useSetAtom, and useAtomValue hooks. In react-context-slices you define either React Context or Redux slices, while in jotai you define atoms. The need for the migration from the first to the second was a high memory usage by react-context-slices when using React Context slices.

Atom-based: splits states into tiny pieces of data called atoms, which can be written to and read from using React hooks. In this group, we have Recoil and Jotai.

jotai Is the signal based state manager I recommended, offering the best developer experience (in my opinion) as it eliminates the necessity to define and update a global store

Jotai 🧙‍♂️

Atomic libs include Recoil, Jotai, and the one I maintain: Zedux. Zedux especially was designed to work well with sockets and RxJS observables and has been getting some traction recently, so of course I recommend checking it out. Feel free to hmu with any questions.

Global State Machine: Jotai (great for any state that needs to be stored globally for your application)

```

It looks like it’s mutating, but both the reducers and update() uses immer* under the hood, so we still respect immutability under the hood.

Cami supports redux devtools so you can use that for time-travel debugging too!

—-

* https://github.com/immerjs/immer

I heard that immutability libraries like immer.js [0] help with this. Anyone go this way and had good success? Is this 'the way'?

[0]: https://immerjs.github.io/immer/

So, that's what we built, and we built it in the most obvious way — using JavaScript Proxy objects to track mutations and reflect those changes across Appsmith’s framework. Initially things looked good — it worked, aside from a few hacks to make some data types work with map and set, and we were following the example of other projects that had similar requirements. If it was good enough for them, it should be good enough for us, right?

But is immer really the ultimate answer? The performance problem of immer is more prominent in large arrays and deep-level object scenarios. See this issue description, many authors in the community began to try to make breakthroughs, and noticed that structura and mutative, I found that it is indeed many times faster than immer as they said, but it still fails to solve the problem of both fast speed and good development experience. I will analyze the two issues in detail below.

I like immer for this kind of thing: https://github.com/immerjs/immer

It gives you immutable updates without getting bogged down in FP abstractions.

I've always been a huge fan of immer for these case. For your code, it would simply turn into setGridData((prev) => produce(prev, draft => applyChanges(changes, draft)) but I recommend you go over their documentation to fully understand how it works

If you're trying to make things actually immutable, Object.freeze and deep copies can clutter things up pretty good, have you considered using something like immer? (https://immerjs.github.io/immer/)

If you want to set up from Context, Zustand is your best bet. It offers an extremely simple API that lets you create a store with values and functions. Then, you can access that store from anywhere in your application to read and write values. Reactivity included! If you want to store nested object data in your store, consider using Immer alongside Zustand to easily change nested state.