Error boundaries do the same for component trees. They allow you to isolate errors and send them to an error reporting service.
Let's learn they are used to catch errors thrown by promises.
Instructor: [00:00] We're using a dynamic import and Suspense to lazily load this PokemonDetail component. It's pretty dang cool. Unfortunately, not all promises resolve. Some get rejected. It could be the fault of a network error or a server error, but things are bound to go sideways, eventually.
[00:24] How can we catch any errors and gracefully tell the user that something went wrong? Let's start by making an error and see what happens. Let's duplicate this line real quick and instead of importing here, let's just return a rejected promise. Save, and this is what our customers see. Nothing.
[00:48] If one thing is true about React, is that the errors are fantastic. Let's open up a console to see what might be hiding here for us. Go to the errors and we see three errors.
[01:00] First, the uncaught exception that made the page blow up, and second, a really great error. The above error occurred in one of your React components. They provide a stack trace for us. App, div, Suspense and the unknown component.
[01:17] Here's the advice. Consider adding an error boundary to your tree to customize error handling behavior. Visit this special URL to learn more about error boundaries. I love it when they sneak a URL into the console.
[01:33] On this doc, you'll find everything that you need to know about error boundaries, the philosophy, practices, and how to customize them. Truth be told, you might not really need them. You can just copy and paste this example and go from there. I've got it, let's put it in our code.
[01:52] One thing you will have to change is logging to the error service. In a production app, you should have one of these, but for our purposes, we can just use console.log or, better yet, console.error.
[02:07] All that's left is to see how to implement it. We just need to wrap the error boundary component around a part of our tree that may be throwing an error. Let's start by wrapping it around the entire component tree.
[02:25] Go back to our app, and we'll notice a few changes. First, we get this new fancy error, which we are displaying from our error boundary component. Second, we see that the callstack is still provided to us, but without the URL, because now we are using the error boundary. Third, we see the error that would have been provided to our error logging service, if we weren't using console.error.
[02:50] Now that we're catching and handling the error, there are a few things that we can do to improve the user experience.
[02:57] First off, this error boundary doesn't need to be at the top of the component. If we move it closer to our React Suspense code, reformat and save, we'll see that we can prevent other parts of the application from being interrupted, just because we got an error here. That's pretty great. It protects the rest of our application from anything that might go wrong in one component tree.
[03:23] We're going to also customize this to take some type of customizable fallback, just like our Suspense component. I'll say something pity like, "Couldn't catch 'em all." Jump in to our error boundary code, cut this and use this.props.fallback.
[03:47] If we want to make sure that something gets presented regardless of whether or not we put a fallback in, we can use static, default props and define our fallback. Something went wrong.
[04:04] As I've said before, you really only need one of these for an entire application, so we can put this in another file for import elsewhere. Let's open up our tree view, create a new file, error boundary, and paste.
[04:21] We'll have to make two changes. We'll need to export defaults for our class component, and, because of the JSX, we'll need to [inaudible] React.
[04:37] Back in our app, we can now import our error boundary from our local path.
[04:46] You can lazy load those components with confidence knowing that no matter what happens in your component, you can catch it, present a fallback, and protect the rest of your application.