Bug: Rules of Hooks should not allow passing a Hook as a function argument

[gaearon]

See failing test in #22044.

Even if the PR gets stale, we should implement this check because currently it lets you circumvent the rule.

[byronluk]

Can I pick this up?

[gaearon]

There’s already PR started with failing tests in #22044 so let’s let its author respond first?

[byronluk]

There’s already PR started with failing tests in #22044 so let’s let its author respond first?

Of course

[otakustay]

This rule should allow passing hooks to hooks like useWrappedState(useRecoilState)) which can ensure non-conditional call to hooks

[gaearon]

Note the difference between foo(useX) (not ok) and foo(useX()) (ok).

[kmercer1]

To clarify, would useWrappedState(useRecoilState)) be not okay as it is passing a hook as a function argument? What about in the scenario that useWrappedState will alway call that hook?

Does this change when passing a hook to a component via a prop?

[eisenpony]

This rule sounds too broad. Why shouldn't hooks be passed as parameters to other hooks?
I asked something similar in #23230 ..

I can understand why foo(useX) is not okay, but useFoo(useX) should be fine, as long as the rules apply to useFoo.
Likewise, I think <Foo useX={useY} /> (currently passes lint) or even <Foo useX={() => useY()} /> (currently fails lint) is okay.

[vileven]

Hello everyone,

I want to bring attention to a question raised by @eisenpony, as it's not entirely clear to me whether passing hooks as props is considered incorrect practice.

It appears that this method is widely used in the industry, and in my view, it doesn't seem to violate the rules of hooks. However, the existence of open issues regarding this matter is quite concerning to me.

just an example of usage this approach by Wolt: https://careers.wolt.com/en/blog/engineering/injecting-hooks-into-react-components