Tag: tsconfig.json

Learn TypeScript #2, The Basics of How Types Work Pt. 2

Transcript:

Alright! Hello everybody! This video is a continuation of the previous video on basic types in TypeScript. Today we’re going to go over null, undefined, using types with functions, and a few other things. If you missed the previous video, you might want to go back and watch it before you start this one.

There was one thing I wanted to clarify from the last video. I mentioned that I strongly recommended against using the “any” type unless you absolutely had to, and I wanted to explain why. The basic reason, is that when you use the “any” type, you’re opting out of nearly all the checks that TypeScript is going to do for you, so if you’re going to use “any” everywhere, you might as well not use TypeScript.

I know when I first started with TypeScript, I often found myself reaching for the “any” type, but 99 times out of 100 it wasn’t because I needed it, but because I was being too lazy to stop and think about what the type of something should be.

If you take those few seconds to stop and think, you’ll end up with a program that’s better designed, you’ll have code that’s easier to come back to and change later, and you’ll have TypeScript helping you catch errors later on down the road.

Enough about “any”, let’s talk about null and undefined. This is definitely one of the more confusing parts of JavaScript (and therefore TypeScript), so let’s talk a little bit about what null and undefined are.

You may be familiar with null from other programming languages, and the easiest way I find to think about null is that it represents the explicit lack of any value. To use our player object example, if there’s no player 2, it’s possible that we’d want our player2 variable to be null.

Undefined is very similar, as it also represents the lack of any value. The key difference here, is that any variable you never assign any value to is going to hold undefined by default (though you could also explicitly set a variable to undefined, like you can set one to null).

Generally, I see null as representing a value that is intentionally missing for whatever reason. Whereas when undefined crops up in a program, it’s often a mistake or accident rather than intentional, since it implys that somebody created a variable, but then never used it. Due to this, I recommend that if you need to indicate a lack of some value, use null rather than undefined, since it will be clearer to people interacting with your code that this was probably something intentional.

All that aside, how do we use null and undefined in our TypeScript programs?

You’ll remember from our last video that we could use “null” as a type just like string or boolean or anything else, and the same is true for undefined.

Notice that if we create a variable and don’t assign to it before we try to use it, TypeScript helpfully gives us an error.

Note also that even though variables default to undefined, we can’t actually assign the value undefined or null without specifying it as part of the type of our variable.

You’ll see we get the error “type undefined is not assignable to type string”.

We have to explicitly list it as a possible type in order for TypeScript to accept it.

You’ll see also that if we try to use this variable without checking wether the type is actually not null or undefined, TypeScript will give us an error as well, preventing us from making a common mistake where we assume a variable has been assigned some value, when really it could just as easily be null or undefined.

Now, it’s important to note that TypeScript actually has a mode where it will be less strict about the usage of null and undefined. If we open up tsconfig.json and scroll down a little bit, you’ll see that I have strict mode enabled. This enables all of the strict checks listed below it, but the one that is most relevant to us is “strictNullChecks”.

If we disable strictNullChecks, or if we disable all of the stricter checks, then TypeScript will become more lenient on us and allow us to set any variable to null or undefined.

Generally I recommend against taking the more lenient route, because it leads to a lot of potential errors where you assume something will have a proper value when it actually doesn’t. It’s much better to explicitly document using the type system that a value might be null or undefined so that the proper checks can be made.

The main case for disabling strictNullChecks is if you’re converting a large existing JavaScript project over to TypeScript and you don’t want to have to deal with all the errors TypeScript starts throwing about not checking for null or undefined, since it’s common for JavaScript developers to be more wonton about their use of null and undefined. You can still get a lot of benefits out of TypeScript without enabling strictNullChecks, but again, I recommend that you enable it.

Alright, enough of that. So far we’ve been defining a number of top-level variables, but your functions will need to have types as well in order to indicate the types of arguments they accept, and what values they return.

As an example, let’s create a super simple function that tests whether a given number is a magical number.

We’ll say that if the number we pass in is 5, then it’s a magical number, otherwise it’s not magical.

And you’ll see TypeScript is complaining at me here, saying “parameter num implicitly has type any”, which is a fancy way of saying I haven’t told it what type num is, so it is forced to assume that it could be anything.

We can remedy this using syntax identical to what we used above. After the variable name, I’ll simply put a colon and the type of the variable, which is number, and now TypeScript is happy with us again.

Not only is TypeScript happy with us, but if we go to use this function, our text editor can use the definition to display what the expected types of the function arguments are.

Another thing you can do is explictly specify what the return type of your function should be. We do this in nearly the same way as declaring variables, by placing a colon after the closing parenthesis and then the type.

So, for example, if we were like hey we should have been returning a boolean, we can specify that, and then TypeScript will let us know anywhere that we are returning the wrong value.

In general as I have said before, TypeScript is going to be really good at inferring what the types of things are. So, here we could have gotten rid of the return type and TypeScript would’ve correctly guessed that boolean was what we wanted.

However, I recommend you always type your arguments in your return values for functions, that way it’s really clear to you and your team members that the function is operating in the way you intended it to operate. And it’s really easy to just glance at the line of code where you create the function and get an idea of what you expect it to do.

One other type I wanted to mention that is related to functions is called the never type.

It’s an peculiar type that means a function will never return. The two main examples of this are functions that you call to throw an error, or functions that create an infinite loop.

For example, something you’ll see in many code bases is is a function named “fail” or “panic” or something similar. This is the function you call when things have gone horribly wrong and you need to log an error and end the program.

Since the program dies before the function fully finishes, the never type is very appropriate. It’s never going to return. You can indicate that this is the case for a particular function by specifying “never” as the return type.

Likewise, if you have a function containing an infinite loop, that will also have a return value of never. The loop never stops, so the function will never return.

Keep in mind that “never” is distinct from a function that returns no value at all. For example, many times when you’re working with graphics for a game or something else, you’ll have some code that updates those graphics on-screen. There isn’t much information this code needs to return anywhere, so it’ll successfully successfully draw the graphics, and then return nothing. For that case, TypeScript has a “void” type, which you can use to indicate that the function will return at some point, but it’s not going to return any value.

That’s it for how to use basic types, and that wraps up this video.

If you’ve enjoyed what you learned here, you might like the TypeScript course that I’m working on. Head on over to https://typescriptbyexample.com, and you can put your email address in a the bottom of the page to be notified when it’s ready.

Thanks so much for watching! See you in the next video.

Learn TypeScript #1, The Basics of How Types Work Pt. 1

Transcript:

Alright! Hello everybody! Today I wanted to go over the very basics of how types work in TypeScript so that you can start using them in the programs you write. If you don’t have TypeScript installed, please see my prior videos where I’ll show you how to do that.

I’ve started here with an empty folder in VSCode, feel free to use any text editor of your choice as you follow along. Before we get started, we’re going to open up a terminal and create a TypeScript configuration file by running “tsc –init”

We’ll change one line for now, we’ll uncomment “lib” and add “dom” and “es5”. This will let us reference objects that exist when running JavaScript in a web browser, which I’ll be using for examples later.

Now let’s create our main.ts file where we’ll be doing everything today.

And then we’re going to go to Terminal and click Run Task and do “tsc –watch” to get TypeScript compiling our files.

We’re going to start off by specifying types on simple variables. There are three ways to declare variables in TypeScript and JavaScript: var, let, and const. In terms of specifying types, they all work the same, so for this video I’ll simply be using “let”. In a future video I’ll cover the differences between var, let, and const, but like I said, for this video, it doesn’t matter.

The first thing I want to touch on before we go much further is that TypeScript is really good at figuring out what the types of things are supposed to be without you needing to explicitly declare it.

For example, you can see here in Visual Studio Code we can figure out what the type of something is by hovering our mouse over, and we see that it pops up what it knows about this variable. It says “bar: number” which means it’s already figured out that this variable is a number!

However, you’ll often want to explicitly tell TypeScript what the type of something is, so that’s what we’ll be focusing on in this video. Now, as it turns out, that same syntax VS Code showed is what you’ll use to explicitly tell TypeScript what the type of something is.

We can say here “let bar: number”, and TypeScript will start enforcing that this variable can only ever be a number.

You’ll see if we try to change the value to something else, we’ll get a red underline, an error listed below, that this string is not assignable to type number.

And, just to be clear, if we remove the “: number”, we still get the same error, because TypeScript is going to assume that once you assign a variable a value of one type, that it was probably a mistake to use any other type.

Let’s start with the what the most basic types are in TypeScript: booleans, numbers, and strings.

A boolean can be either true or false.

A number can be any number you like.

And a string can be any text that you like.

Building on top of the basic types, we can do arrays of things by specifying the type and then writing an opening and closing square bracket.

There is an alternate syntax for writing arrays, though generally I don’t see it used as often. Instead of using the square brackets, you write Array, an angle bracket, the type contained in the array, and then a closing angle bracket. This is completely the same as the other syntax used above, and like I said, this is used much less commonly so I say default to the first way unless you have a good reason to do otherwise.

There’s also a variation on arrays called a tuple. If you’re not familiar with the term, you can think of it simply as an array with a fixed number of elements.

Say there are two pieces of data you always want to store next to each other, but don’t want to use a full object for whatever reason, you could do that with a tuple. For example, we could store student test scores as a tuple containing their name and the score.

Here, Tom has a score of 97.5.

And and you can see this is expecting just a string and a number. If we tried to add something extra to this, it’s going to complain to us that we’ve created something that’s not the size that we said it would be or if we try to remove a required element TypeScript will also complain. And similarly if we use the wrong type it’ll complain as well.

To be clear, you can have an array of tuples by combining the same syntaxes we’ve just learned. So we’ll have our tuple type followed by the opening and closing square brackets.

And then we would fill it out like so, and TypeScript will enforce that all the types we use here are correct.

Declaring our own objects is done in a very similar way to arrays. Let’s say we’re making a game, and we have a player with an x position, a y position, and some health. How would we do this in TypeScript?

We’ll start by declaring our player variable, then doing a colon and some curly brackets, and inside the curly brackets we can put the names and types of the fields we want to create. You’ll see that this looks very much like the actual object value that we’ll assign to the player object.

If we accidentally put in the wrong type, or if we add a field that doesn’t exist as part of the definition, TypeScript will complain at us.

It’ll often be the case that you’ll want to re-use object definitions like this in multiple places, say if you had multiple players or various functions that manipulated player objects.

There are two ways of doing this, the first and most direct way is to use the “type” keyword to define a new name that is equal to the type we defined for players earlier.

Then we can use that name in place of the full object definition for the player.

Another option with very similar effects is to create an interface. You can do this by using the “interface” keyword followed by the name you want, and then describe the types of the fields for the object like we did earlier. We can use it as the type for our player in exactly the same way that we used the definition created by the “type” keyword.

When is it appropriate to use “type” instead of “interface”? The answer is a bit too long for this video, so stay tuned for a future one where I’ll cover that. For now, just use whichever one you feel most comfortable with.

Now, going back to our example, you can see that I can now create a second player using the same type.

And if I update the definition of the type, TypeScript will helpfully give us an error on both of them, indicating that they need to be updated to match the changes.

This feedback loop is super common in development. You’ll often notice some object needs a new field, or needs that field altered in some way. You can make that change in the central type definition, and then go through and fix each of the places where TypeScript indicates that there’s a problem. No need to worry about whether you updated every spot or not, TypeScript will let you know.

Now is a good time to mention that in the same that way we created and used our own types above, we can use existing types that are part of JavaScript or JavaScript’s environment. For example, because we included “dom” in our “lib” in tsconfig.json, we can create a variable and declare that it’s type is HTMLElement. We could then fill this variable with something from the DOM API, like document.getElementById.

Now, you’ll see here that TypeScript is complaining about our HTMLElement variable. Why is it doing that? Well, if we hover over document.getElementById, we’ll see that the type is actually “HTMLElement | null”, the pipe character there means “or”. That’s because if we specify an ID that doesn’t exist on the page, this function will return null.

Using the pipe character to join multiple types together creates what’s called a “union type”, which is simply a fancy way of saying “this variable could be a few different things”. There’s a lot more to union types, and other related features in TypeScript that we’ll cover in a future video, for now, just be aware that if you have a variable that could be a couple different types, you can list those types by separating them with the pipe character.

Okay, so that covers the basics of builtin types and defining our own types. There’s a couple of catch-all types that I want to show you.

The first one is an “object”, and that represents anything that isn’t a basic type. So it’ll be anything that isn’t a boolean or a number or a string. For example, it could be HTMLElement or our Player type, or anything else. This is a type that I don’t see used all that often, because generally you either know what the object type is, or you use the even more general “any” keyword.

“any” is exactly what it sounds like. If you say the type of your variable is “any”, you’re saything that it can hold any value possible. Generally if you’re using TypeScript, I strongly recommend against reaching for “any” unless you absolutely have to use it. Though if you’re interacting with non-TypeScript code, it can be appropriate to use “any” in order to deal with JavaScript libraries that truly can produce any value.

Next up is talking about how to handle null, undefined, and how to use types with function definitions, but this video is already running a little long, so stay tuned for part 2 of this video.

If you’re enjoying what you’re learning here, you might like the TypeScript course that I’m working on. Head on over to https://typescriptbyexample.com, and you can put your email address in the form at the bottom of the page to be notified when it’s ready.

Thank you so much for watching. I’ll see you in the next video.