Open ziriax opened 3 years ago
I feel like a case of convergent evolution. I just learned Recoil exists this weekend. I have been using my own poor man's versions of useRecoilState
, useRecoilStateLoadable
, and essentially a useRecoilStateLazyLoadable
(which Recoil doesn't seem to have as of yet). I love Loadable
. My version of Loadable
can have an additional "uncalled"
state.
I like how the API of Recoil read-ably formalizes things I've only been following as patterns. Recoil also adds waitForAll()
, waitForAny()
, and waitForNone()
(all of which I haven't even thought of writing).
Anyway, I've only ever read and haven't written any RxJS-like-things, but feel like I know enough to answer your question. There is a danger that if you separate hasValue
, hasError
, and loading
states into separate cases that the rendered HTML outputs will have different dimensions and that can cause glitchy-ness. That can be solved by fixing loading skeleton and error output dimensions. Or it can be solved by combining separate states in a way similar to this example from an "article" I wrote.
I believe that Recoil does not suffer from the problem described in the article you linked. It would be a great thing to have tests for!
Hi, I would love to contribute to this issue. Although would need some directions where I could or should start writing tests for. I've been using recoil for about 3 months now, and would love to contribute. I think this would be a great place to start.
@davidmccabe
Thank you in advance. :)
I'm not sure If this is still an ongoing issue, but if so I would love to work on unit tests.
Sure, we would definitely welcome more tests!
First of all thanks for sharing this amazing library!
I was wondering, is Recoil glitch-free?
Many other reactive frameworks - like RxJS - don't have a synchronous dataflow architecture, so updates happen in the order of subscription, and not per se in the correct dependency order (like FlapJAX, SodiumFRP, Reactive-Banana, etc).
I'm curious if Recoil solves this problem?