immer

t across thunks & reducers, and is hard to follow. This also needs refactoring across every dispatch site. B. Use something like redux sagas to intercept actions: We found this to be "too" flexible and felt that we were better-off without the entire machinery of spawning sagas on the fly. We wanted it to be easy to reason about what happens when an action is dispatched looking at the code without having to debug what all sagas could be running at that particular point of time. C. Move the action handling higher up: requires extensively refactoring the reducers. The solution we settled on was redux-loop [1]: A port of elm's effect system to redux. This was neat because we could easily convert the reducer to return loops instead of states, and thereby easily get access to full state and dispatch while retaining the ability to follow through the complete action handling flow from a single starting point that didn't change. TypeScript support in redux-toolkit is also kind'a bolted on and users are recommended different approaches when they care about type-safety. It proved to be a pain to communicate junior devs multiple times that you should use leave the reducers as empty object and instead use "extraReducers" with builder API. We found using immer[2] (for managing immutable state) and unionize[3] (for handling discriminated union of action types) directly to be a much better solution than redux-toolkit's abstractions. [1]: https://github.com/redux-loop/redux-loop [2]: https://github.com/immerjs/immer [3]: https://www.npmjs.com/package/unionize

ly experimenting with rollback netcode. A key part of rollback is saving your state every frame so you can load it if you need to roll back, and ECS makes saving/loading easier to reason about, since you can "just" grab the components and their associated entity IDs and load if needed (EnTT, for the record, has an API for this[1], though they leave the actual (de)serialization up to you). Of course, JS's lack of a memcpy equivalent makes this much harder than what you could do in C++, which has lead me to experiment with immer[2] in my ECS, which uses structural sharing to avoid mutation, so you can get a "copy" of your state by just keeping a reference to it, as future updates will make new objects. This, of course, theoretically could make a ton of garbage (e.g. updating your position every frame would create a new Position object every frame), which is not great for high performance games. I'm not sure how bad this will be in practice - JS GC is relatively smart and fast these days, but I haven't tried doing much beyond little pong or platformer demos yet. I'd also imagine that, like, doing a deep clone of my state tree every frame (or doing the whole JSON.stringify/parse dance if I stick to primitive values) probably generates just as much garbage. Maybe if I could integrate immer with some kind of object pool it'd avoid these issues, but I have no idea how useful object pooling is in practice in JS... [1] https://github.com/skypjack/entt/wiki/Crash-Course:-entity-c... [2] https://github.com/immerjs/immer