You need those because it is hard to do the following:
- client JavaScript sends a request to server
- server sends back data
- client updates what the user sees
This is hard to do notably because when the update happens, several things might need to change on the webpage at the same time.
Notably, new elements might need to be added to the webpage, which in turn means that new bindings such as button clicks have to be added to those, in a way that keeps the page working.
The only way to do this basically is to have a functional dependency graph that keeps everything in the page in working state as updates come.
Website: reactjs.org
React officially recommends that you use Next.js[ref], so just do it. It just sets up obvious missing functionality from raw React.
React feels like a good. But it also feels impossible to use/learn sometimes.
Its main design goal is to reduce DOM changes to improve rendering times.
And an important side effect of that is that it becomes easier to do stuff of the type:and then the new comment easily gets the callback attached to it.
- user creates a new comment that appears on screen without page reload
- comment has a delete button, which is JavaScript callback activated
And it also ends up naturally doubling as a template engine.
But React can also be extremely hard to use. It can be very hard to know what you can and cannot do sometimes, then you have to stop and try to understand how react works things better:The biggest problem is that it is hard to automatically detect such errors, but perhaps this is the same for other frontend stuff. Though when doing server-side rendering, the setup should really tell you about such errors, so you don't just discover them in production later on.
- cannot update a component while rendering a different component warning in React
- Rendered more hooks than during the previous render.
- cannot use hooks from helpers:
Is is also very difficult to understand precisely why hooks run a certain number of times.
Examples under: react.
- react/hello.html
- react/hello-func.html: Hello World with a React function component instead of classes. At page load console shows:and then after each click:
Mainso we understand thatonClick MainMaininsanely functions both as the constructor and as the render function in React function components. - react/hello-func-use-callback.html: same as react/hello-func.html but with useCallback. TODO no advantages in this case? When does it help?
- react/hello-without-jsx.html: Hello World in pure JavaScript, without JSX. Exactly equivalent to react/hello.html. Documented at: reactjs.org/docs/react-without-jsx.html Understanding this is fundamental to understanding React.
- react/prop-change.html: shows what gets called as parameters flow down through the tree.By looking at the console, we see all
renderget called every time, even ifpropsdidn't change, but not the constructors.After page load the console contains:Main.constructor Main.render NotMain.constructor NotMain.render NotMain2.constructor NotMain2.renderThen, every time we click the button it adds:handleClick Main.render NotMain.render NotMain2.renderNote how thepropsofNotMainonly change every other click, butrenderstill gets called every time.In order to makeReactnot re-render when there are not changes, you have to either:- define the
shouldComponentUpdatemethod of class components - wrap functional components in
React.memo
- define the
- react/prop-change-hook.html: same as react/prop-change.html, but using hooks. The notable difference is that functional components don't have a clear constructor/render separation, the function just gets called every time. Then React does some magic to ensure that
useStatereturns the current state, except for the first render where they return the initial value. - react/prop-change-hook-use-memo.html: TODO forgot if this example is useful, was tring to use
useMemo - react/prop-change-child.html: shows what child prop changes do not call render on parent,
Maindoes not show up on console when you click underNotMain - react/hook-from-function-fail.html: TODO got some errors that seemed linked to this on a larger program, but failed to minimize them here
- react/hook-different-number-of-times.html: this illustrates one of the cardinal points of using hooks: you must always call them the same number of times, otherwise it fails with:In the case of
React has detected a change in the order of Hooks called by Main. This will lead to bugs and errors if not fixed.
useState, we can kind of understand why this happens: React must use the order of calls to determine which state variable to return at each point in time. - react/hello-hook-use-effect.html: just checking when it gets called. Happens after every render
handleClick Main useEffect useEffect2 - TODO create a test
\a[react/img-broken.html]
How React works bibliography:
- www.netlify.com/blog/2019/03/11/deep-dive-how-do-react-hooks-really-work/ shows how
uesStateworks under the hood with crazy closures - medium.com/@gethylgeorge/how-virtual-dom-and-diffing-works-in-react-6fc805f9f84e
Minimal reproduction: stackoverflow.com/questions/62336340/cannot-update-a-component-while-rendering-a-different-component-warning/70317831#70317831
Minimal React hello world example. As you click:By opening a web inspector, you can see that only modified elements get updated. So we understand that JSX parses its "HTML-like" into a tree, and then propagates updates on that tree.
- one counter increments every time
- the other increments every two clicks
By looking at the terminal, we see that
render() does get called every time the button is clicked, so the tree of elements does get recreated every time. But then React diffes thing out and only updates things in the DOM where needed.Failed attempt at minimizing nodejs/next/ref-twice outside of Next.js.
Dummy example of using a React
ref This example is useless and to the end user seems functionally equivalent to react/hello.html.It does however serve as a good example of what react does that is useful: it provides a "clear" separation between state and render code (which becomes once again much less clear in React function components.
Notably, this example is insane because at:
we are extracing state from some random HTML string rather than having a clean JavaScript variable containing that value.
<button onClick={() => {
elem.innerHTML = (parseInt(elem.innerHTML) + 1).toString()In this case we managed to get away with it, but this is in general not easy/possible.
React function components do produce shorter code. But they are also impossible to understand without knowing what is their corresponding class component.
Hooks were introduced much after classes, and just require less code, so everyone is using them now instead of classes.
This should only be used for things that happen outside of the state that React trackes, e.g.
window event handlers.Examples:
Framework built on top of React.
Officially recommended by React[ref]:
Recommended ToolchainsIf you’re building a server-rendered website with Node.js, try Next.js.
gothinkster/realworld blog example by Ciro Santilli: node Express Sequelize Next.js realworld example app.
Basically what this does is to get server-side rendering just working by React, including hydration, which is a good thing.
Next.js sends the first pre-rendered HTML page along with the JavaScript code. Then, JavaScript page switches just load the API data.
Next.js does this nicely by forcing you to provide page data in a serialized JSON format, even when rendering server-side (e.g. the return value of
getServerSideProps). This way, it is also able to provide either the full HTML, or just the JSON.Some general downsides:
- it does feel like they don't document deployment very well however, especially non-Vercel options, which is the company behind Next.js. I'm unable to find how to use a non Vercel CDN with ISR supposing that is possible.
- Next.js is very opinionated, and like any opinionated library it is sometimes hard to know why something is/isn't happening, and sometimes it is hard/impossible to do what you want with it unless they add support. They have done good progress, but even as of 2022, some aspects just feel so immature, some major-looking use cases are not very well done.
In theory, Next.js could be the "ultimate frontend framework". It does have a lot of development difficulties that need to be ironed out, but the general concepts, and things it tries to integrate, including e.g. webpack, TypeScript, etc. are good. Maybe the question is when will someone put it together with an amazing backend library and dominate and finally put an end to the infinite number of Js Frameworks!
In-tree examples at: github.com/vercel/next.js/tree/canary/examples
In order to offer its amazing features, Next.js is also extremely opinionated, which means that if something wasn't designed to be possible, it basically isn't.
No prerender with custom server? It forces you to write your API with next as well? Or does it mean something else?
TODO can it statically generate pages that are created at runtime? E.g. if I create a new blog post, will it automatically upload a static page? It seems that yes, and that this is exactly what Incremental Static Regeneration means:However, Ciro can't find any mention of how to specify where the pages are uploaded to... this is pat of the non-Vercel deployment problem.
- github.com/vercel/next.js/discussions/25410
- vercel.com/docs/next.js/incremental-static-regeneration
- github.com/vercel/next.js/discussions/17711
- www.reddit.com/r/nextjs/comments/mvvhym/a_complete_guide_to_incremental_static/
- github.com/vercel/next.js/discussions/11552#discussioncomment-115595
- stackoverflow.com/questions/62105756/how-to-use-aws-with-next-js
- github.com/vercel/next.js/discussions/17080
- github.com/vercel/next.js/discussions/16852
Can't ISR prerenter by URL query parameters:
That plus the requirement to have one page per file under
pages/ leads to a lot of useless duplication, because then you are forced to place the URL parameters on the pathnames."Module not found: Can't resolve 'fs'" Hell. The main reason this happens seems to be the that in a higher order component, webpack can't determine if callbacks use the require or not to remove it from frontend code. Fully investigated and solved at:
Overviews:
- www.reddit.com/r/reactjs/comments/8evy5d/what_are_the_downsides_to_nextjs/ 2017 What are the downsides to Next.js?
TODO answer:
Our examples are located under nodejs/next:
- nodejs/next/hello-world: a hello world. There's an in-tree one at: github.com/vercel/next.js/tree/e75361fd03872b097e817634c049b3185f24cf56/examples/hello-world, but ours is truly minimal
- nodejs/next/hoc: shows how to use a higher order component (HOC) to factor out
getStaticPropsacross two pages: nodejs/next/hoc/pages/index.js and nodejs/next/hoc/pages/notindex.js - nodejs/next/typescript: simple TypeScript example, minimized from: github.com/vercel/next.js/tree/d61b0761efae09bd9cb1201ff134ed8950d9deca/examples/with-typescriptNotably, that shows how
requireerrors are avoided in that case as mentioned at: stackoverflow.com/questions/64926174/module-not-found-cant-resolve-fs-in-next-js-application/70363153#70363153 - nodejs/next/localStorage: a counter that is persistent across page reloads by using
localStorage. Used in: stackoverflow.com/questions/54819721/next-js-access-localstorage-before-rendering-page/68136224#68136224
Solved ones:
- solved by preview mode in Next.js 12:
- ISR was basically unusable for CRUD websites because you can't force a one-off immediate page update:
The goal of this example is to understand when states and effects happen when changing between different routes that use the same component.
Behavior is follows:
- visit: localhost:3001/1
- click
count++. This makescount: 1 - click "2" to visit localhost:3001/2
- outcome: count is still 1
This is likely because in React the state kept in the virtual DOM structure, and identical structure implies identical state. So when we change from post 1 to 2, we still have a
Post object, and state is unchanged.Next if we click:then the count is back to 0. This is because we changed the
- "Index" to go to localhost:3000
- "1" to go to localhost:3001/1
Post object in the DOM to Index and back, which resets everything.This example also illustrates how to prevent this from happening with
useEffect.Bibliography:
This is a minimal reproducible example for the terrible problem of external effects applying twice to refs for effects that are not idempotent and thus blowup if applied twice.
The issue is currently discussed at: react.dev/learn/synchronizing-with-effects#step-3-add-cleanup-if-needed (archive) which says "you need to cleanup the thing yourself". web.archive.org/web/20240720100401/https://react.dev/learn/synchronizing-with-effects#subscribing-to-events is also says that for the specific case of
addEventListener.But that's annoying! Can't we just somehow tell if we applied twice or not to avoid having to implement a cleanup? What if a third party system does not provide a cleanup at all?
Is the correct solution to just just have a
useEffect with empty dependency list? Seems to be good according to posts and to ESLint!Tried to do a React only reproduction at: react/ref-twice.html.
Bibliography:
Ciro Santilli's questions:
In this example we attempt to inject React elements into statically rendered HTML coming from the server, and properly hydrate them.
Questions by Ciro Santilli:
Bibliography:
Articles by others on the same topic
There are currently no matching articles.