Is it possible to Kotlin Compiler to infer type parameter based on another type parameter?
That is: infere type A based on type B in IntExample class definition.
interface Something<A>
interface Example<A, B: Something<A>>
class IntSomething: Something<Int>
// Can Int type parameter be infered by the fact that IntSomething is Something<Int>?
// If yes, how could I specify B type parameter without specifing the infered A (Int) parameter?
class IntExample: Example<Int, IntSomething> // specifing Int here is painful!!!
Imagine that we have more type parameters like that - it will be a lot of boilerplate to specify each of them if some might (theoretically) be infered.
EDIT
After an exhausive respose from #KarstenGabriel I will extend the previous example, to make it clear what for are the type parameters used here:
interface Something<A> {
val sth: A
}
interface Example<A, B: Something<A>> {
val eg: A
val something: B
}
data class IntSomething(override val sth: Int): Something<Int>
data class DescribedIntSomething(
override val sth: Int,
val description: String
): Something<Int>
data class DescribedIntExample(
override val eg: Int,
override val something: DescribedIntSomething,
): Example<Int, DescribedIntSomething> // specifing Int here is painful
fun main() {
val describedIntExample = DescribedIntExample(
eg = 1,
something = DescribedIntSomething(1, "Just one")
)
// We have to know that something is DescribedIntSomething to read description.
// Neither `*` nor `Something<Int>` is sufficient, we need `B: Something<Int>` to keep B
val description = describedIntExample.something.description
println(description)
}
So we use the type parameters - A and B as return values from eg and something
Wildcard * cannot be used as it is just means Any?. We need to keep concrete type B. (eg. to enable reading descrption in the example)
Wildcard Something<Int> cannot be used instead of B: Something<Int> for thesame reason as in the point 1
Maybe it's just an language design level. The type parameters do exists in compile time (they are ereased later), so theoretically Example<DescribedIntSomething> might be sufficient instead of Example<Int, DescribedIntSomething> as DescribedIntSomething is Something<Int>
The problem is solvable under some assumptions:
You need type parameter A only for property eg (or other similar cases).
It is okay that we convert eg to be a function instead of a property.
The desired value for eg (your A value) can be derived from the value of something (your B value). If the value is not inferred, then it is not possible to infer the type.
We then define Example without parameter A:
interface Example<B: Something<*>> {
val something: B
}
Now inside Example we do not have a possibility to obtain the type parameter of B's Something-type, because the type parameter A on Something<A> does not exist at runtime. So you cannot just ask an instance of B what its Something-type parameter is. And Kotlin cannot do that either, because the information is just not there, if it is not explicitly specified by you.
But we know the parameter at compile-time which can be used in an inline extension function with reified parameter:
inline fun <reified A, reified B : Something<A>> Example<B>.eg(): A = something.sth
(documentation: Inline functions with reified parameter and Extension functions)
In this example I assume that eg should have the value of sth.
If you want eg to be private, you can put the function inside Example and make it private, otherwise it must be specified outside of the class.
Now you can just define your subclass without the need to specify the former A parameter, but you still have the right type:
data class DescribedIntExample(
override val something: DescribedIntSomething,
) : Example<DescribedIntSomething>
fun main() {
val describedIntExample = DescribedIntExample(
something = DescribedIntSomething(1, "Just one")
)
val x = describedIntExample.eg() // x has inferred type Int and value 1
}
val (_,time) = time { Thread.sleep(1000) }
I see the Left Hand Side (LHS) has a val, so its declaring a variable. LHS also has some kind of a function syntax which does not look like a lambda declaration. What is (_,time)? Don’t you have to give a type to the time on the LHS? I understand the RHS perfectly well: it is a function which accepts a lambda as parameter and is named ‘time’. Original code
The left hand side is called destructuring.
If you try to assign an instance of a data class (or any class which has componentN functions) to a variable you can desturcture it. This means that you can assign its internals to variables. The _ syntax indicates that you don't care about the first item.
Example:
class Foo(val first: String, val second: String) {
operator fun component1() = first
operator fun component2() = second
}
Usage:
val (first, second) = Foo("first", "second")
If you use data classes you don't need to create the componentN functions, they are generated for you.
Equivalent data class:
data class Foo(val first: String, val second: String)
Found this in kotlin documentation about function and lambda
class IntTransformer: (Int) -> Int {
override operator fun invoke(x: Int): Int = TODO()
}
val intFunction: (Int) -> Int = IntTransformer()
In this page, it says that you can implement function type to class like an interface. How does it work? Can you give me some explanation every each part and give me an example how this is done?
From what I understand, IntTransformer expand/implement anonymous function that takes int as argument and output type, but I still didn't know how does it work...
Thanks
You can think of a function type sort of like an interface that has a single function named invoke with the parameters and return type matching its definition.
So
(Int) -> String
is very much like
interface Foo {
operator fun invoke(param: Int): String
}
And so if a class inherits from (Int) -> String, you would do it in exactly the same way as you would to inherit Foo above. You could say the function inheritance is more versatile, because it allows your class to be passed around as a functional argument directly instead of having to get a reference to its invoke function using ::invoke.
In Kotlin doc, type check use is but when I write this code
var a="hello"
if (a is String) print(a)
There is a warning
warning: check for instance is always 'true'
if (a is String) print(a)
^
Thank you very much for all answers.
In your example, "hello" is a String literal. In Kotlin, even when you omit the type for a variable, its type is inferred. The compiler infers the type for var a from the initializer expression, and so the type of a is String. The warning you are getting means that the expression a that you check is always a String.
Your variable declaration is equivalent to var a: String = "hello", i.e. the variable may only reference a String, assigning any other type is not allowed.
For example, if you change the variable declaration to var a: Any = "hello", there will be no warning since the variable now may hold an instance of any type, not just a String.
I just figured out how to use type check by learning from Swift
open class fruit {}
class banana: fruit() {}
fun test( a: fruit ) {
if (a is banana) print("ok")
}
test(banana())
How is it related to extension functions? Why is with a function, not a keyword?
There appears to be no explicit documentation for this topic, only the assumption of knowledge in reference to extensions.
It is true that there appears to be little existing documentation for the concept of receivers (only a small side note related to extension functions), which is surprising given:
their existence springing out of extension functions;
their role in building a DSL using said extension functions;
the existence of a standard library function with, which given no knowledge of receivers might look like a keyword;
a completely separate syntax for function types.
All these topics have documentation, but nothing goes in-depth on receivers.
First:
What's a receiver?
Any block of code in Kotlin may have a type (or even multiple types) as a receiver, making functions and properties of the receiver available in that block of code without qualifying it.
Imagine a block of code like this:
{ toLong() }
Doesn't make much sense, right? In fact, assigning this to a function type of (Int) -> Long - where Int is the (only) parameter, and the return type is Long - would rightfully result in a compilation error. You can fix this by simply qualifying the function call with the implicit single parameter it. However, for DSL building, this will cause a bunch of issues:
Nested blocks of DSL will have their upper layers shadowed:
html { it.body { // how to access extensions of html here? } ... }
This may not cause issues for a HTML DSL, but may for other use cases.
It can litter the code with it calls, especially for lambdas that use their parameter (soon to be receiver) a lot.
This is where receivers come into play.
By assigning this block of code to a function type that has Int as a receiver (not as a parameter!), the code suddenly compiles:
val intToLong: Int.() -> Long = { toLong() }
Whats going on here?
A little side note
This topic assumes familiarity with function types, but a little side note for receivers is needed.
Function types can also have one receiver, by prefixing it with the type and a dot. Examples:
Int.() -> Long // taking an integer as receiver producing a long
String.(Long) -> String // taking a string as receiver and long as parameter producing a string
GUI.() -> Unit // taking an GUI and producing nothing
Such function types have their parameter list prefixed with the receiver type.
Resolving code with receivers
It is actually incredibly easy to understand how blocks of code with receivers are handled:
Imagine that, similar to extension functions, the block of code is evaluated inside the class of the receiver type. this effectively becomes amended by the receiver type.
For our earlier example, val intToLong: Int.() -> Long = { toLong() } , it effectively results in the block of code being evaluated in a different context, as if it was placed in a function inside Int. Here's a different example using handcrafted types that showcases this better:
class Bar
class Foo {
fun transformToBar(): Bar = TODO()
}
val myBlockOfCodeWithReceiverFoo: (Foo).() -> Bar = { transformToBar() }
effectively becomes (in the mind, not code wise - you cannot actually extend classes on the JVM):
class Bar
class Foo {
fun transformToBar(): Bar = TODO()
fun myBlockOfCode(): Bar { return transformToBar() }
}
val myBlockOfCodeWithReceiverFoo: (Foo) -> Bar = { it.myBlockOfCode() }
Notice how inside of a class, we don't need to use this to access transformToBar - the same thing happens in a block with a receiver.
It just so happens that the documentation on this also explains how to use an outermost receiver if the current block of code has two receivers, via a qualified this.
Wait, multiple receivers?
Yes. A block of code can have multiple receivers, but this currently has no expression in the type system. The only way to achieve this is via multiple higher-order functions that take a single receiver function type. Example:
class Foo
class Bar
fun Foo.functionInFoo(): Unit = TODO()
fun Bar.functionInBar(): Unit = TODO()
inline fun higherOrderFunctionTakingFoo(body: (Foo).() -> Unit) = body(Foo())
inline fun higherOrderFunctionTakingBar(body: (Bar).() -> Unit) = body(Bar())
fun example() {
higherOrderFunctionTakingFoo {
higherOrderFunctionTakingBar {
functionInFoo()
functionInBar()
}
}
}
Do note that if this feature of the Kotlin language seems inappropriate for your DSL, #DslMarker is your friend!
Conclusion
Why does all of this matter? With this knowledge:
you now understand why you can write toLong() in an extension function on a number, instead of having to reference the number somehow. Maybe your extension function shouldn't be an extension?
You can build a DSL for your favorite markup language, maybe help parsing the one or other (who needs regular expressions?!).
You understand why with, a standard library function and not a keyword, exists - the act of amending the scope of a block of code to save on redundant typing is so common, the language designers put it right in the standard library.
(maybe) you learned a bit about function types on the offshoot.
When you call:
"Hello, World!".length()
the string "Hello, World!" whose length you're trying to get is called the receiver.
More generally, any time you write someObject.someFunction(), with a . between the object and the function name, the object is acting as the receiver for the function. This isn't special to Kotlin, and is common to many programming languages that use objects. So the concept of a receiver is likely very familiar to you, even if you haven't heard the term before.
It's called a receiver because you can think of the function call as sending a request which the object will receive.
Not all functions have a receiver. For example, Kotlin's println() function is a top-level function. When you write:
println("Hello, World!")
you don't have to put any object (or .) before the function call. There's no receiver because the println() function doesn't live inside an object.
On the receiving end
Now let's look at what a function call looks like from the point of view of the receiver itself. Imagine we've written a class that displays a simple greeting message:
class Greeter(val name: String) {
fun displayGreeting() {
println("Hello, ${this.name}!")
}
}
To call displayGreeting(), we first create an instance of Greeter, then we can use that object as a receiver to call the function:
val aliceGreeter = Greeter("Alice")
val bobGreeter = Greeter("Bob")
aliceGreeter.displayGreeting() // prints "Hello, Alice!"
bobGreeter.displayGreeting() // prints "Hello, Bob!"
How does the displayGreeting function know which name to display each time? The answer is the keyword this, which always refers to the current receiver.
When we call aliceGreeter.displayGreeting(), the receiver is aliceGreeter, so this.name points to "Alice".
When we call bobGreeter.displayGreeting(), the receiver is bobGreeter, so this.name points to "Bob".
Implicit receivers
Most of the time, there's actually no need to write this. We can replace this.name with just name and it will implicitly point to the name property of the current receiver.
class Greeter(val name: String) {
fun displayGreeting() {
println("Hello, $name!")
}
}
Notice how that differs from accessing a property from outside the class. To print the name from outside, we'd have to write out the full name of the receiver:
println("Hello, ${aliceGreeter.name}")
By writing the function inside the class, we can omit the receiver completely, making the whole thing much shorter. The call to name still has a receiver, we just didn't have to write it out. We can say that we accessed the name property using an implicit receiver.
Member functions of a class often need to access many other functions and properties of their own class, so implicit receivers are very useful. They shorten the code and can make it easier to read and write.
How do receivers relate to extensions?
So far, it seems like a receiver is doing two things for us:
Sending a function call to a specific object, because the function lives inside that object
Allowing a function convenient and and concise access to the other properties and functions that live inside the same object
What if we want to write a function that can use an implicit receiver for convenient access to the properties and functions of an object, but we don't want to (or can't) write our new function inside that object/class? This is where Kotlin's extension functions come in.
fun Greeter.displayAnotherGreeting() {
println("Hello again, $name!")
}
This function doesn't live inside Greeter, but it accesses Greeter as if it was a receiver. Notice the receiver type before the function name, which tells us that this is an extension function. In the body of the extension function, we can once again access name without its receiver, even though we're not actually inside the Greeter class.
You could say that this isn't a "real" receiver, because we're not actually sending the function call to an object. The function lives outside the object. We're just using the syntax and appearance of a receiver because it makes for convenient and concise code. We can call this an extension receiver, to distinguish it from the dispatch receiver that exists for functions that are really inside an object.
Extension functions are called in the same way as member functions, with a receiver object before the function name.
val aliceGreeter = Greeter("Alice")
aliceGreeter.displayAnotherGreeting() // prints "Hello again, Alice!"
Because the function is always called with an object in the receiver position before the function name, it can access that object using the keyword this. Like a member function, an extension function can also leave out this and access the receiver's other properties and functions using the current receiver instance as the implicit receiver.
One of the main reasons extension functions are useful is that the current extension receiver instance can be used as an implicit receiver inside the body of the function.
What does with do?
So far we've seen two ways to make something available as an implicit receiver:
Create a function inside the receiver class
Create an extension function outside the class
Both approaches require creating a function. Can we have the convenience of an implicit receiver without declaring a new function at all?
The answer is to call with:
with(aliceGreeter) {
println("Hello again, $name!")
}
Inside the block body of the call to with(aliceGreeter) { ... }, aliceGreeter is available as an implicit receiver and we can once again access name without its receiver.
So how come with can be implemented as a function, rather than a language feature? How is it possible to simply take an object and magic it into an implicit receiver?
The answer lies with lambda functions. Let's consider our displayAnotherGreeting extension function again. We declared it as a function, but we could instead write it as a lambda:
val displayAnotherGreeting: Greeter.() -> Unit = {
println("Hello again, $name!")
}
We can still call aliceGreeter.displayAnotherGreeting() the same as before, and the code inside the function is the same, complete with implicit receiver. Our extension function has become a lambda with receiver. Note the way the Greeter.() -> Unit function type is written, with the extension receiver Greeter listed before the (empty) parameter list ().
Now, watch what happens when we pass this lambda function as an argument to another function:
fun runLambda(greeter: Greeter, lambda: Greeter.() -> Unit) {
greeter.lambda()
}
The first argument is the object that we want to use as the receiver. The second argument is the lambda function we want to run. All runLambda does is to call the provided lambda parameter, using the greeter parameter as the lambda's receiver.
Substituting the code from our displayAnotherGreeting lambda function into the second argument, we can call runLambda like this:
runLambda(aliceGreeter) {
println("Hello again, $name!")
}
And just like that, we've turned aliceGreeter into an implicit receiver. Kotlin's with function is simply a generic version of this that works with any type.
Recap
When you call someObject.someFunction(), someObject is acting as the receiver that receives the function call
Inside someFunction, someObject is "in scope" as the current receiver instance, and can be accessed as this
When a receiver is in scope, you can leave out the word this and access its properties and functions using an implicit receiver
Extension functions let you benefit from the receiver syntax and implicit receivers without actually dispatching a function call to an object
Kotlin's with function uses a lambda with receiver to make receivers available anywhere, not just inside member functions and extension functions
Kotlin knows the concept of a function literals with receiver. It enables access on visible methods and properties of a receiver of a lambda within its body without having to use any additional qualifier. That's very similar to extension functions in which you can as well access members of the receiver object inside the extension.
A simple example, also one of the greatest functions in the Kotlin standard library, is apply:
public inline fun <T> T.apply(block: T.() -> Unit): T {
block()
return this
}
Here, block is a function literal with receiver. This block parameter is executed by the function and the receiver of apply, T, is returned to the caller. In action this looks as follows:
val foo: Bar = Bar().apply {
color = RED
text = "Foo"
}
We instantiate an object of Bar and call apply on it. The instance of Bar becomes the receiver of apply. The block, passed as an argument in curly brackets does not need to use additional qualifiers to access and modify the properties color and text.
The concept of lambdas with receiver is also the most important feature for writing DSLs with Kotlin.
var greet: String.() -> Unit = { println("Hello $this") }
this defines a variable of type String.() -> Unit, which tells you
String is the receiver
() -> Unit is the function type
Like F. George mentioned above, all methods of this receiver can be called in the method body.
So, in our example, this is used to print the String. The function can be invoked by writing...
greet("Fitzgerald") // result is "Hello Fitzgerald"
the above code snippet was taken from Kotlin Function Literals with Receiver – Quick Introduction by Simon Wirtz.
Simply put ( without any extra words or complications) , the "Receiver" is the type being extended in the extension function or the class name. Using the examples given in answers above
fun Foo.functionInFoo(): Unit = TODO()
Type "Foo" is the "Receiver"
var greet: String.() -> Unit = { println("Hello $this") }
Type "String" is the "Receiver"
Additional tip: Look out for the Class before the fullstop(.) in the "fun" (function) declaration
fun receiver_class.function_name() {
//...
}
Simply put:
the receiver type is the type an extension function extends
the receiver object is the object an extension function is called on; the this keyword inside the function body corresponds to the receiver object
An extension function example:
// `Int` is the receiver type
// `this` is the receiver object
fun Int.squareDouble() = toLong() * this
// a receiver object `8` of type `Int` is passed to the `square` function
val result = 8.square()
A function literal example, which is pretty much the same:
// `Int` is the receiver type
// `this` is the receiver object
val square: Int.() -> Long = { toLong() * this }
// a receiver object `8` of type `Int` is passed to the `square` function
val result1 = 8.square()
val result2 = square(8) // this call is equal to the previous one
The object instance before the . is the receiver. This is in essence the "Scope" you will define this lambda within. This is all you need to know, really, because the functions and properties(varibles, companions e.t.c) you will be using in the lambda will be those provided within this scope.
class Music(){
var track:String=""
fun printTrack():Unit{
println(track)
}
}
//Music class is the receiver of this function, in other words, the lambda can be piled after a Music class just like its extension function Since Music is an instance, refer to it by 'this', refer to lambda parameters by 'it', like always
val track_name:Music.(String)->Unit={track=it;printTrack()}
/*Create an Instance of Music and immediately call its function received by the name 'track_name', and exclusively available to instances of this class*/
Music().track_name("Still Breathing")
//Output
Still Breathing
You define this variable with and all the parameters and return types it will have but among all the constructs defined, only the object instance can call the var, just like it would an extension function and supply to it its constructs, hence "receiving" it.
A receiver would hence be loosely defined as an object for which an extension function is defined using the idiomatic style of lambdas.
Typically in Java or Kotlin you have methods or functions with input parameters of type T. In Kotlin you can also have extension functions that receive a value of type T.
If you have a function that accepts a String parameter for example:
fun hasWhitespace(line: String): Boolean {
for (ch in line) if (ch.isWhitespace()) return true
return false
}
converting the parameter to a receiver (which you can do automatically with IntelliJ):
fun String.hasWhitespace(): Boolean {
for (ch in this) if (ch.isWhitespace()) return true
return false
}
we now have an extension function that receives a String and we can access the value with this