https://overreacted.io/making-setinterval-declarative-with-react-hooks/
Dan Abramov article explains well how to work with hooks, state, and the setInterval() type of API!
Dan Abramov! Is one of the React maintaining team! So known and I personally love him!
Quick explanation
The problem is the problem of how to access the state with a useEffect()
that executes only once (first render)!
Note: For a deep explanation of why the state isn't updated within the useEffect callback
and other inside useEffect callbacks
. Check the last section about closures and react re-render ...
The short answer is: by the use of refs (useRef)! And another useEffect() that run again when update is necessary! Or at each render!
Let me explain! And check the Dan Abramov solution! And you'll get better the statement above at the end! With a second example that is not about setInterval()!
=>
useEffect() either run once only, or run in each render! or when the dependency update (when provided)!
Accessing state can be possible only through a useEffect() that run and render each relevant time!
Or through setState((state/*here the state*/) => <newStateExpression>)
But if you want to access the state inside useEffect() => re-run is necessary! meaning passing and executing the new callback!
That doesn't work well with setInterval! If you clear it and re-set it each time! the counter get reset! Leading to no execution if the component is re-rendering fast! And make no sense!
If you render only once! The state is not updated! As the first run, run a one callback! And make a closure! The state is fixed! useEffect(() => { <run once, state will stay the same> setInterval(() => { <state fixed as closure of that time> }) }, [])
.
For all such kind of situation! We need to use useRef
! (refs)!
Save to it a callback that holds the state! From a useEffect()
that rerenders each time! Or by saving the state value itself in the ref
! Depending on the usage!
Dan abramov solution for setInterval (simple and clean)
That's what you are looking for!
useInteval hook (by Dan Abramov)
import React, { useState, useEffect, useRef } from 'react';
function useInterval(callback, delay) {
const savedCallback = useRef();
// Remember the latest callback.
useEffect(() => {
savedCallback.current = callback;
}, [callback]);
// Set up the interval.
useEffect(() => {
function tick() {
savedCallback.current();
}
if (delay !== null) {
let id = setInterval(tick, delay);
return () => clearInterval(id);
}
}, [delay]);
}
Usage
import React, { useState, useEffect, useRef } from 'react';
function Counter() {
let [count, setCount] = useState(0);
useInterval(() => {
// Your custom logic here
setCount(count + 1);
}, 1000);
return <h1>{count}</h1>;
}
We can see how he kept saving the new callback at each re-render! A callback that contains the new state!
To use! it's a clean simple hook! That's a beauty!
Make sure to read Dan article! As he explained and tackled a lot of things!
setState()
Dan Abramov mentioned this in his article!
If we need to set the state! Within a setInteral! One can use simply setState() with the callback version!
useState(() => {
setInterval(() => {
setState((state/*we have the latest state*/) => {
// read and use state
return <newStateExpression>;
})
}, 1000);
}, []) // run only once
we can even use that! Even when we are not setting a state! Possible! Not good though! We just return the same state value!
setState((state) => {
// Run right away!
// access latest state
return state; // same value (state didn't change)
});
However this will make different react internal code part run (1,2,3), And checks! Which ends by bailing out from re-rendering! Just fun to know!
We use this only when we are updating the state! If not! Then we need to use refs!
Another example: useState() with getter version
To showcase the how to work with refs and state access
! Let's go for another example! Here is another pattern! Passing state in callbacks!
import React from 'react';
function useState(defaultVal) {
// getting the state
const [state, setState] = React.useState(defaultValue);
// state holding ref
const stateRef = React.useRef();
stateRef.current = state; // setting directly here!
// Because we need to return things at the end of the hook execution
// not an effect
// getter
function getState() {
// returning the ref (not the state directly)
// So getter can be used any where!
return stateRef.current;
}
return [state, setState, getState];
}
The example is of the same category! But here no effect!
However, we can use the above hook to access the state in the hook simply as below!
const [state, useState, getState] = useState(); // Our version! not react
// ref is already updated by this call
React.useEffect(() => {
setInteval(() => {
const state = getState();
// do what you want with the state!
// it works because of the ref! Get State to return a value to the same ref!
// which is already updated
}, 1000)
}, []); // running only once
For setInterval()
! The good solution is Dan Abramov hook! Making a strong custom hook for a thing is a cool thing to do! This second example is more to showcase the usage and importance of refs, in such state access need or problem!
It's simple! We can always make a custom hook! Use refs! And update the state in ref
! Or a callback that holds the new state! Depending on usage! We set the ref on the render (directly in the custom hook [the block execute in render()])! Or in a useEffect()
! That re-run at each render or depending on the dependencies!
Note about useEffect() and refs setting
To note about useEffect()
useEffect => useEffect runs asynchronously, and after a render is painted on the screen.
- You cause a render somehow (change state, or the parent re-renders)
- React renders your component (calls it)
- The screen is visually updated
- THEN
useEffect
runs
Very important of a thing! useEffect()
runs after render() finishes and the screen is visually updated! It runs last! You should be aware!
Generally, however! Effects should be run on useEffect()! And so any custom hook will be ok! As its useEffect()
will run after painting and before any other in render useEffect()! If not! As like needing to run something in the render directly! Then you should just pass the state directly! Some people may pass a callback! Imagine some Logic component! And a getState callback was passed to it! Not a good practice!
And if you do something somewhere of some such sense! And talking about ref
! Make sure the refs
are updated right! And before!
But generally, you'll never have a problem! If you do then it's a smell! the way you are trying to go with is high probably not the right good way!
Closure and more. Why the state variables are not having the latest value?
The why you can't simply access the state value directly boils down to the closure notion
. Render function at every re-render go totally with a new call
. Each call has its closures. And at every re-render time, the useEffect callback
is a new anonymous function. With its new scope of value.
And here the dependency array
matter. To access the closure of this call
. And so the recent state of this call. You have to make useEffect
use the new callback
. If dependencies change then that would happen. If not that wouldn't.
And if you do []
then useEffect()
would run the first time only. every new call after the first render. Would always have useEffect
get a new anonymous function
. But none of them is effective or used (run).
The same concept applies to useCallback
and many other hooks
. And all of that is a result of closures
.
callbacks inside the useEffect callback
ex: event listners
, setInterval
, setTimeout
, some(() => {})
, api.run(() => {})
Now even if you update the useEffect callback
through dependency change
. And let's say you did some event listener
or setInterval call
. But you did it conditionally so if already running no more setting it. The setInterval callback
or event listener callback
wouldn't get access to the recent state value. Why? you already guessed. They were created in the first run
, first call space and all the state values are closure passed down at that **time**
. And in the later updates and calls. It's a new render function call
. And a whole different useEffect callback
as well. Sure it's the same function code
. But not same functions at all
. The initial callback of the initial run on that render function call. Is a function that was created back then. And then the event listener callback or the setInterval callback was created back then
. And they are still in memory
and referred to
. The event listener part of the event listener API objects instances, and the object that made the registration and the setInterval part of the node runtime. They had the state closure having the value of the time. And never get to see or know about any other re-render call
. **Unless somehow you. inject something within. That
still referencesor
can access the latest value(references or globals). And those values come from the hooks (useState) and their internal machinery. All it would have is the
closures of the time of the creation.**
Funny metaphor
Most people falling into this trap. Look at the code and saying hey why when the state updates it's not getting the newer value. And the answer is. The state value is something coming from useState, it's not a global variable. And even though you are looking at the same code. The first call and later calls are all different spaces (beings). The only thing making it possible for us to work with functions in this sense is the hooks. And how they store the state and bring it back.
And a good metaphor would be: to get to an office and do some contract and deal. And then later come back and enter the same office, but wait not really the same office rather one that looks the same. But a new office took its place (moved out, moved in, same activity). And you wonder why they didn't recognize you. Yup a bit of an off of a metaphor. But still good a bit.
On the whole I hope that gave you a good sense!