This misses event.shiftKey. All keyboard modifiers disqualify client-side routing.
As for <Link> in its entirety, where you have to use a special component which becomes a link and adds a click handler to it, I think this is generally the wrong solution: it’s better to instead put a single click handler on the entire document to intercept clicks on links. That way, you can just use normal links everywhere and it just works, rather than having to remember to use a whole separate component every time which may or may not be able to pass through the required properties or attributes (e.g. this one only supports class and href). Simpler, smaller and cheaper.
I think React tends to encourage local solutions (using a Link component) over global ones (inserting a global click handler) when possible. Global solutions are less composable and will cause issues if you have multiple independent teams working on the same page (eg maybe they need different click handlers)
In general I would agree, but as soon as you’re using popstate, you’ve broken that completely: it’s inherently global and will conflict with any other compositions. Because of this, the use of a global click handler will make the application more consistent: either all forms of navigation work, or none; whereas if you use <Link> it gives the impression that you can compose, but the back and forward buttons may be broken.
The History API, as it stands, is global state. That makes it one of the very few places where I say it is better to use a global solution, rather than a precise one.
> In general I would agree, but as soon as you’re using popstate, you’ve broken that completely: it’s inherently global and will conflict with any other compositions
It can be problematic to access globals from functional programming contexts, but I don't really see how History would cause conflicts here. Do you have any examples?
If you are really worried about this then the proper way to solve it imo is to pass `history` in as a prop or Context instead of calling it from the global `window` directly. This follows functional paradigms and is more in line with React principles. This has some other advantages as well. For example, if you are using test frameworks like Jest, you can test the link component by mocking the History object being passed in.
The URL is global state. Only one thing can use it to control what’s displayed. It is an error to manipulate `location` or `history` at the widget or library level: they must only be used at the application level.
Passing history in as a property doesn’t solve anything in the application: it’s still global state, and if multiple things try to use it (even if one is trying to use the hash and the other the path), they’ll interfere with each other and you’ll have a bad time.
Passing it as a property may help with mocking-based test frameworks (though I find mocking to normally be a bad technique—you get better results from unit tests and integration tests with real systems), though you may also be able to control the global variables to accomplish the same effect; yet as far as the application is concerned, it’s still global state.
So then: since anything working with the URL is unavoidably working with global state, I see no virtue in avoiding a global event handler.
Global state is never actually necessary. Functional programming doesn't require global state to be turing complete. So I don't see what you mean by it being unavoidable. Passing it in makes it not global state anymore. As I explained in the mocking example. You can run multiple tests at once while initializing a different mock History object to be passed in for every test.
In my experience, generally when people try to use global handlers, it ends up getting bloated as more and more teams add their quirks and features into those handlers. There are numerous ways to mitigate this but I've found that the easiest way to prevent these issues is to simply move the logic into components. And React context gets rid of the need for prop drilling and removes most of the boilerplate involved with passing data to deeply nested components.
The middle mouse button doesn’t trigger a click event, but rather auxclick (like right click also doesn’t, but rather contextmenu and then auxclick if that’s preventDefaulted).
But there are definitely situations where it’s judicious to check event.button, and I missed altKey, too. Here’s the full function Fastmail uses:
Ah indeed, I probably confused it with other mouse events that do not have a button-specific event like `auxclick` and `contextmenu`. Mouse down/up only use the button to differentiate the buttons, which can also explain why that function includes such check (even if unnecessary on the `click` event)
On a side note, I can't believe we still don't have a `visit` or `activate` event that works regardless of hardware, without having to exclude modifier keys.
I don’t think the React root is really any better than the document—you should either go precise, or go global, since popstate is global. If there are some links you don’t want intercepted, they’re as likely to be within the React root as without it. See also my response to woojoo666.
addEventListener("click", function (event) {
const link = event.target.closest("a");
if (
!event.button &&
!event.altKey &&
!event.ctrlKey &&
!event.metaKey &&
!event.shiftKey &&
link &&
link.href.startsWith(location.origin + "/") &&
link.target !== "_blank"
) {
event.preventDefault();
navigate(link.href);
}
}
There are some points of nuance here where you may want to vary (e.g. the determination of eligible hrefs, and handling of href targets), but this is the bones of it. You can also choose to add more functionality to it. Fastmail’s webmail, for example, uses this basic design, but also handles mailto: links specially (taking you to compose a new email), whitelists path patterns (since the domain is used for more than just the webmail app), and one or two other things.
It gets called on all clicks (line 1, it’s a window-level click handler), then finds the link the click target is inside (line 2), and does nothing if it wasn’t inside a link (line 9).
(You might think that it should use .closest("a[href]") instead of .closest("a"), since a:not([href]) is not a link, but in that case, link.href === "", and so it fails the line 10 is-it-eligible-for-client-side-routing test. Note also that except for the “no href attribute” case, HTMLAnchorElement#href gives a full resolved URL, so <a href=""> will produce the document base URL.)
Fundamentally, there’s no such thing as an event handler that triggers only on links—you instead have to use a global event handler that starts by checking whether it’s being triggered on a link.
I do think, though, that woojoo666’s comment and my response are worth bearing in mind. In this specific situation, I think <Link> is not warranted and mildly better avoided; but in a similar situation where global behaviour wasn’t already forced, I would say to keep the separate component, rather than breaking behavioural encapsulation.
Whenever I see someone describe how small something they've written as a replacement for a big library is, I wonder what they've discarded in order to get it so small.
Normally it's backwards compatibility, but in this case it appears to be "everything".
I don't get it...this is a simple library that people can use without bringing in external dependencies. Of course it won't be as functional as React-Router - that goes without saying!!
I would say that a router that doesn't support url patterns, query parameters or state is basically useless for anything other than small static sites that in my personal opinion shouldn't have clientside routing.
This is a nice demonstration of how routers can work for people that are curious, but I would have a lot of concerns about it as production-worthy, especially because you're manually firing browser native events during periods that people generally wouldn't expect them to fire.
I'm saying this as someone that had to write a full clientside routing system for a bespoke web app - routing is very hard and full of edge cases.
If I read this correctly, all the route components are evaluated (i.e. rendered to virtual DOM) and only then one of them is selected and shown (i.e. rendered to HTML DOM). This seems utterly inefficient, even for educational example.
You read that incorrectly. Only the active route component is returned by the Router function, therefore React's render() won't do anything with the others.
There's one thing, you should redirect all the pages to one single endpoint in server side order to use "pushState". Otherwise it will return 404 when you hit the refresh button. If you don't own a server, you can support routing with hashtag "#" and listen to "onhashchange" event instead of "popstate".
Also, if you would like to support nested and dynamic routes (it's not possible with that code snippet in the github repository since it just checks like `path===currentPath`), you might look at the following solution:
Please don’t encourage using a single route and the hash; people should just ensure that suitable document is served from all routable URLs—whether it be a consistent bootstrap file, or (generally preferably) that you do server-side rendering. Using the fragment for routing is only something to do when you can’t do it the proper way, not a tool to reach for from the start.
> There's one thing, you should redirect all the pages to one single endpoint in server side order to use "pushState". Otherwise it will return 404 when you hit the refresh button. If you don't own a server, you can support routing with hashtag "#" and listen to "onhashchange" event instead of "popstate".
Could you explain this one a bit more / maybe some example code I could borrow from? :)
I've used a lot of routers and my favorite is still page.js[1]. It hasn't been updated in years. But it's small, is Express-compatible (i.e. server/client routes can use the same code), and, more importantly, is hackable. I'll never use a router tied to a certain framework again (react, nextjs, etc.) because you trade flexibility for perceived convenience (e.g. using folder structure as route structure, or React component tree as route structure). But it's a terrible trade-off that paints you into a corner later, IMO. Routing can get really niche and site-dependent, so having it fully under your control is worth it.
1kb is likely a lot more than what is shown here; I made a "tiny" but very complete React Router package which is very complete and minified+gzip it's just 1.8kb https://crossroad.page/
The author's goal in this seems to be education, not really optimization or replacement of a bigger framework. And I guess this would probably be much less than 1k!
It’s strange that of the numerous frameworks React seems to have the most churn, it’s had multiple ways of “doing state”, there seems to be a new router every year or at the very least a new major (breaking) version of react router.
Why is routing, of all things, in such need of constant development? I say this as someone that’s made SPAs with Vue a lot and occasionally React.
Because making a simple router or simple state management library from scratch is quite simple. So a lot of people do them to learn how the internals work, like this case here. This makes people better programmers. Some of those folks will then decide to continue the development, put some special things here and there and release it to the world to see if anyone is interested in their version of it. Most of those are ignored, which is fine.
However this makes some people vocally angry, because in theory a developer should know every single packages in the world, but never the internals. Only people of a higher class or caste should be allowed to release stuff to the world. The rest are peasants. How dare they. /s
The reason people say this is because the scale of the problem is much, much worse in front-end javascript than in most other language communites and it's a real problem for people attempting to get to grips with best practice.
If your explanation was correct then we'd see the same problem everywhere at the same scale but there's something about front-end javascript that seems to exacerbate this issue.
I disagree that the scale of the problem is worse in the Frontend.
About being uncharitable: IMO this constant mischaracterization of frontend is the real uncharitable thing here.
I also do a lot of backend work and even the most "traditionally stable" programming languages (or frameworks) have a lot of churn in libraries, frameworks, patterns, toolchains, deployment methods and architectures. Which is "fine", that's how programming works!
But heck, I worked recently on a 8 year old Rails app that had four different methods/libraries/frameworks being used to wrap business code. But in the frontend that would be extremely rare.
If anything the frontend has been quite stable for the last 7 years or so when we settled on mostly React or Vue. The only real big change was Hooks, and even that was incremental and was an attempt to solve real problems in the ecosystem.
This is basically gaslighting. F/E churn is still wild. I’m staring at an article right now titled the top 14 react libraries you must try in 2020. That’s not a normal thing for a ten year old framework.
How is this gaslighting? Just because someone wrote an article to profit on your FOMO doesn't mean React (and also Vue) usage in the real world hasn't been extremely stable over the years. If anything, this kind of article is the one trying to fuck with your psyche.
You don't need any of the libraries in the article to make a good React app/website. Period. No, not even React-Router or Redux. And those two are probably the only ones that you'll see lots of companies/devs that actually ship stuff using. And both of them are not even that complicated, as is demonstrated in this article for the Router. Redux is also quite simple (hence why there's so many alternative implementations).
If anything the ones doing the gaslighting are the ones saying that we devs need to use ultra-complex build systems and gigantic libraries to make a "real app". It's the gaslighting of saying simple architectures "won't ever work" mixed with machismo of "real programmers do X".
Sure, if you want the complexity that comes with using those "must-use" libraries, feel free to have it. I won't say you're wrong by doing so. Maybe you REALLY need all 14 of them, although I doubt it. But please stop assuming that everyone is doing the same or has to. Nor try to deny the reality of me and lots of others getting things done without "14 must use libraries". Because that is the definition of gaslighting.
You seem to be conflating your personal experience here with everything happening around you which as you rightly point out is absolutely filled with this ever moving target that is absolutely impenetrable to beginners.
The React ecosystem looks like the crypto / nft space to outsiders. Nobody seems to agree on anything, no two projects look alike, it’s just this endless loop of people reinventing the same core bits over and over again for the past ten years.
I’m happy for you that this hasn’t been you experience but don’t tell me that the React ecosystem is stable. The reputation of the React ecosystem for many years has famously been that it’s a house of cards.
Nope. I'm not conflating my personal experience, because I'm not denying that other people can have different experiences. All I am saying is that blogs posts created to generate buzz and SEO-driven-traffic are not indicative of what the ecosystem really is, let alone of what an individual experience can be.
You, on the other hand seems to be claiming that the only reality that matters is the one that matches your own biases about an ecosystem that yourself seems to only participate from the periphery of SEO-spam blog posts.
It's funny that you're started by calling me a gaslighter when it is exactly what you're doing.
Claiming that React looks like Crypto/NFT only makes sense from anyone taking advice from such blogs.
The core bits have definitely not been reinvented, as the library itself hasn't really changed much to the outside in eight years other than the adoption of hooks and a few incremental advances.
Please spare the condescending "I’m happy for you" part. You're clearly not part of the industry that is using React in a stable way, and you don't seem to want to be part of. Even if your description of the React ecosystem were true, you're clearly on a personal vendetta here and has no intention to see anything change for the better.
> we'd see the same problem everywhere at the same scale
we do. Every C app implements its own linked list or hash table library. The entire Scheme community is nothing but toy interpreters of various stages of completeness (you can tell a project is serious when they implement call/cc). How many game engines do you think exist? It's a meme that game devs like to spend more time on their pet game engine than actually making their game. How many ORMs do you think exist for <language-of-choice>? At least half a dozen. At least. For any given language. Python, Ruby, Go[1]. ORMs, in particular, seem to get created over and over again. Probably because they are trivial to implement and allows one to voice their opinions on SQL abstraction (bike shedding).
The scale of the problem on FE is because if you want to work in the browser, you need to use JS. There's a diversity of engineers working on the FE and with them comes a diversity of ideas. People have different use cases in mind and build packages to satisfy those use cases.
There's nothing special about JS, but there is something special about being forced to use JS.
It really depends on what you mean by saying wheels are being "reinvented". Is it really wheel reinvention, or is it a small incremental evolution that fixes one pain point, but has to keep all other aspects the same?
In this same thread there is a comment about how this little code snippet "throws away backward compatibility, plus everything else".
New tech has to have backwards compatibility to even be taken serious by most people. But backwards compatibility requires not only reinventing the wheel, but also having it being on the exact same size and material as before so it fits where it was used.
The only way to satisfy both of the groups above is to evolve tech that already exists and implement new findings in already existing libraries, but that will also enrage a lot of people. An example: React Hooks.
What a ridiculous hyperbolic comment in response to a question around why one particular library has a ton more development than any others in the same space.
React isn't a framework. In order to build an app, you have to cobble together a bunch of other libraries to use with it to build something meaningful.
As someone coming from an Angular shop, I started a new job a few months ago that uses React on the frontend. My main gripe so far is there is no one "way" to do X in a react-based app. For an enterprise with lots of developers, Angular seems like a better fit - unless you want to spend months coming up with standards. Angular has its own issues of course - it's much more rigid and has a steeper learning curve.
Are people still having this strawman argument over if react is a library or a framework? Would my argument have changed if I'd put "framework/library" in my comment? No.
I find React perfectly fine for writing SPAs, as long as what is being written is actually suited to being an SPA. Otherwise I find it as dreadful as any time you use a hammer to pound a screw in place.
React is a view library. It deliberately focused on doing one thing, and doing it well -- declaratively painting the view layer. It's not comparable to more general/monolithic frontend development solutions like Angular/Vue/etc. Those tools have far different scopes. So of course React doesn't prescribe routing, state management, project styling solutions, build bundling, or any other frequently lamented "missing" features. They're only omissions to folks misunderstanding React's use case.
Now there are solutions comparable to Vue/etc which use React -- Next.js, Create React App, etc. Comparisons, tradeoffs, and criticisms are totally fair there. Just don't blame React for churn in third party implementations of features outside of its purview.
As a side note, I love that React is so narrowly scoped. It lets the community build on, experiment with, and coalesce around amazing solutions to other big features outside of React's scope. That Darwinian approach has created broad front end solutions that to me are a joy to use compared to similarly featured frameworks like Angular.
It doesn’t actually do it well and has been totally out of step with how modern browsers work for years now. It’s old and poorly made technology. The fact that it chose to do one thing and it can’t even get that right says a lot about it.
Sadly, for reasons that remain a mystery to me it’s one of those things like Apple or crypto where some of its loudest fans have decided to make it a part of their identity rather than a tool they enjoy using.
React actually does more than "one thing", and it does those things quite well, actually: First, it provides a relatively performant diffing mechanism that gives a happy medium between manual diffing and templating; it also provides reusable components that are easy to make, easy to maintain and are able to easily hold local data and have an ergonomic API; finally it also provides a sane of of passing data between layers.
Other frameworks have adopted some of those concepts in different ways, and it works quite well. Other technologies have other ways of solving the same problems, and they also work! Of course, it is absolutely not perfect and doesn't solve "all problems", but that's how mature technologies generally are.
Modern browsers have tried to provide a competing technology with WebComponents, but despite being native to the platform and embraced en masse not only by browser makers but by large companies, it hasn't taken off like React. Please take a time to reflect as to why something that is "there" isn't as widely used as something that requires a 50kb library to even start. And no, it's not a conspiracy, nor it is marketing. There are very good technical reasons for that.
About the "fans" part: please look at the mirror. You seem to be projecting a lot here. People use React and other similar libraries because it gets the job done. A lot of your posts seem to be about competing technologies, or are anti-Apple, or anti-something. Criticism is perfectly fine, but you're the one making hating something part of your identity and the projection is blurring your ability to empathise with what others are saying.
> Why is routing, of all things, in such need of constant development?
Because requirements change. In the last few years there was a lot of churn in React due to the introduction of async rendering and server-side rendering, and React routers were strongly affected by both of those changes.
This misses event.shiftKey. All keyboard modifiers disqualify client-side routing.
As for <Link> in its entirety, where you have to use a special component which becomes a link and adds a click handler to it, I think this is generally the wrong solution: it’s better to instead put a single click handler on the entire document to intercept clicks on links. That way, you can just use normal links everywhere and it just works, rather than having to remember to use a whole separate component every time which may or may not be able to pass through the required properties or attributes (e.g. this one only supports class and href). Simpler, smaller and cheaper.