Edit Page

Generics

As in Java, classes in Kotlin may have type parameters:

class Box<T>(t: T) {
  var value = t
}

In general, to create an instance of such a class, we need to provide the type arguments:

val box: Box<Int> = Box<Int>(1)

But if the parameters may be inferred, e.g. from the constructor arguments or by some other means, one is allowed to omit the type arguments:

val box = Box(1) // 1 has type Int, so the compiler figures out that we are talking about Box<Int>

Variance

One of the most tricky parts of Java’s type system is wildcard types (see Java Generics FAQ). And Kotlin doesn’t have any. Instead, it has two other things: declaration-site variance and type projections.

First, let’s think about why Java needs those mysterious wildcards. The problem is explained in Effective Java, Item 28: Use bounded wildcards to increase API flexibility. First, generic types in Java are invariant, meaning that List<String> is not a subtype of List<Object>. Why so? If List was not invariant, it would have been no better than Java’s arrays, since the following code would have compiled and caused an exception at runtime:

// Java
List<String> strs = new ArrayList<String>();
List<Object> objs = strs; // !!! The cause of the upcoming problem sits here. Java prohibits this!
objs.add(1); // Here we put an Integer into a list of Strings
String s = strs.get(0); // !!! ClassCastException: Cannot cast Integer to String

So, Java prohibits such things in order to guarantee run-time safety. But this has some implications. For example, consider the addAll() method from Collection interface. What’s the signature of this method? Intuitively, we’d put it this way:

// Java
interface Collection<E> ... {
  void addAll(Collection<E> items);
}

But then, we would not be able to do the following simple thing (which is perfectly safe):

// Java
void copyAll(Collection<Object> to, Collection<String> from) {
  to.addAll(from); // !!! Would not compile with the naive declaration of addAll:
                   //       Collection<String> is not a subtype of Collection<Object>
}

(In Java, we learned this lesson the hard way, see Effective Java, Item 25: Prefer lists to arrays)

That’s why the actual signature of addAll() is the following:

// Java
interface Collection<E> ... {
  void addAll(Collection<? extends E> items);
}

The wildcard type argument ? extends T indicates that this method accepts a collection of objects of some subtype of T, not T itself. This means that we can safely read T’s from items (elements of this collection are instances of a subclass of T), but cannot write to it since we do not know what objects comply to that unknown subtype of T. In return for this limitation, we have the desired behaviour: Collection<String> is a subtype of Collection<? extends Object>. In “clever words”, the wildcard with an extends-bound (upper bound) makes the type covariant.

The key to understanding why this trick works is rather simple: if you can only take items from a collection, then using a collection of Strings and reading Objects from it is fine. Conversely, if you can only put items into the collection, it’s OK to take a collection of Objects and put Strings into it: in Java we have List<? super String> a supertype of List<Object>.

The latter is called contravariance, and you can only call methods that take String as an argument on List<? super String> (e.g., you can call add(String) or set(int, String)), while if you call something that returns T in List<T>, you don’t get a String, but an Object.

Joshua Bloch calls those objects you only read from Producers, and those you only write to Consumers. He recommends: “For maximum flexibility, use wildcard types on input parameters that represent producers or consumers”, and proposes the following mnemonic:

PECS stands for Producer-Extends, Consumer-Super.

NOTE: if you use a producer-object, say, List<? extends Foo>, you are not allowed to call add() or set() on this object, but this does not mean that this object is immutable: for example, nothing prevents you from calling clear() to remove all items from the list, since clear() does not take any parameters at all. The only thing guaranteed by wildcards (or other types of variance) is type safety. Immutability is a completely different story.

Declaration-site variance

Suppose we have a generic interface Source<T> that does not have any methods that take T as a parameter, only methods that return T:

// Java
interface Source<T> {
  T nextT();
}

Then, it would be perfectly safe to store a reference to an instance of Source<String> in a variable of type Source<Object> – there are no consumer-methods to call. But Java does not know this, and still prohibits it:

// Java
void demo(Source<String> strs) {
  Source<Object> objects = strs; // !!! Not allowed in Java
  // ...
}

To fix this, we have to declare objects of type Source<? extends Object>, which is sort of meaningless, because we can call all the same methods on such a variable as before, so there’s no value added by the more complex type. But the compiler does not know that.

In Kotlin, there is a way to explain this sort of thing to the compiler. This is called declaration-site variance: we can annotate the type parameter T of Source to make sure that it is only returned (produced) from members of Source<T>, and never consumed. To do this we provide the out modifier:

abstract class Source<out T> {
  abstract fun nextT(): T
}

fun demo(strs: Source<String>) {
  val objects: Source<Any> = strs // This is OK, since T is an out-parameter
  // ...
}

The general rule is: when a type parameter T of a class C is declared out, it may occur only in out-position in the members of C, but in return C<Base> can safely be a supertype of C<Derived>.

In “clever words” they say that the class C is covariant in the parameter T, or that T is a covariant type parameter. You can think of C as being a producer of T’s, and NOT a consumer of T’s.

The out modifier is called a variance annotation, and since it is provided at the type parameter declaration site, we talk about declaration-site variance. This is in contrast with Java’s use-site variance where wildcards in the type usages make the types covariant.

In addition to out, Kotlin provides a complementary variance annotation: in. It makes a type parameter contravariant: it can only be consumed and never produced. A good example of a contravariant class is Comparable:

abstract class Comparable<in T> {
  abstract fun compareTo(other: T): Int
}

fun demo(x: Comparable<Number>) {
  x.compareTo(1.0) // 1.0 has type Double, which is a subtype of Number
  // Thus, we can assign x to a variable of type Comparable<Double>
  val y: Comparable<Double> = x // OK!
}

We believe that the words in and out are self-explaining (as they were successfully used in C# for quite some time already), thus the mnemonic mentioned above is not really needed, and one can rephrase it for a higher purpose:

The Existential Transformation: Consumer in, Producer out! :-)

Type projections

Use-site variance: Type projections

It is very convenient to declare a type parameter T as out and have no trouble with subtyping on the use site. Yes, it is, when the class in question can actually be restricted to only return T’s, but what if it can’t? A good example of this is Array:

class Array<T>(val size: Int) {
  fun get(index: Int): T { /* ... */ }
  fun set(index: Int, value: T) { /* ... */ }
}

This class cannot be either co- or contravariant in T. And this imposes certain inflexibilities. Consider the following function:

fun copy(from: Array<Any>, to: Array<Any>) {
  assert(from.size == to.size)
  for (i in from.indices)
    to[i] = from[i]
}

This function is supposed to copy items from one array to another. Let’s try to apply it in practice:

val ints: Array<Int> = arrayOf(1, 2, 3)
val any = Array<Any>(3)
copy(ints, any) // Error: expects (Array<Any>, Array<Any>)

Here we run into the same familiar problem: Array<T> is invariant in T, thus neither of Array<Int> and Array<Any> is a subtype of the other. Why? Again, because copy might be doing bad things, i.e. it might attempt to write, say, a String to from, and if we actually passed an array of Int there, a ClassCastException would have been thrown sometime later.

Then, the only thing we want to ensure is that copy() does not do any bad things. We want to prohibit it from writing to from, and we can:

fun copy(from: Array<out Any>, to: Array<Any>) {
 // ...
}

What has happened here is called type projection: we said that from is not simply an array, but a restricted (projected) one: we can only call those methods that return the type parameter T, in this case it means that we can only call get(). This is our approach to use-site variance, and corresponds to Java’s Array<? extends Object>, but in a slightly simpler way.

You can project a type with in as well:

fun fill(dest: Array<in String>, value: String) {
  // ...
}

Array<in String> corresponds to Java’s Array<? super String>, i.e. you can pass an array of CharSequence or an array of Object to the fill() function.

Star-projections

Sometimes you want to say that you know nothing about the type argument, but still want to use it in a safe way. The safe way here is to define such a projection of the generic type, that every concrete instantiation of that generic type would be a subtype of that projection.

Kotlin provides so called star-projection syntax for this:

  • For Foo<out T>, where T is a covariant type parameter with the upper bound TUpper, Foo<*> is equivalent to Foo<out TUpper>. It means that when the T is unknown you can safely read values of TUpper from Foo<*>.
  • For Foo<in T>, where T is a contravariant type parameter, Foo<*> is equivalent to Foo<in Nothing>. It means there is nothing you can write to Foo<*> in a safe way when T is unknown.
  • For Foo<T>, where T is an invariant type parameter with the upper bound TUpper, Foo<*> is equivalent to Foo<out TUpper> for reading values and to Foo<in Nothing> for writing values.

If a generic type has several type parameters each of them can be projected independently. For example, if the type is declared as interface Function<in T, out U> we can imagine the following star-projections:

  • Function<*, String> means Function<in Nothing, String>;
  • Function<Int, *> means Function<Int, out Any?>;
  • Function<*, *> means Function<in Nothing, out Any?>.

Note: star-projections are very much like Java’s raw types, but safe.

Generic functions

Not only classes can have type parameters. Functions can, too. Type parameters are placed before the name of the function:

fun <T> singletonList(item: T): List<T> {
  // ...
}

fun <T> T.basicToString() : String {  // extension function
  // ...
}

To call a generic function, specify the type arguments at the call site after the name of the function:

val l = singletonList<Int>(1)

Generic constraints

The set of all possible types that can be substituted for a given type parameter may be restricted by generic constraints.

Upper bounds

The most common type of constraint is an upper bound that corresponds to Java’s extends keyword:

fun <T : Comparable<T>> sort(list: List<T>) {
  // ...
}

The type specified after a colon is the upper bound: only a subtype of Comparable<T> may be substituted for T. For example

sort(listOf(1, 2, 3)) // OK. Int is a subtype of Comparable<Int>
sort(listOf(HashMap<Int, String>())) // Error: HashMap<Int, String> is not a subtype of Comparable<HashMap<Int, String>>

The default upper bound (if none specified) is Any?. Only one upper bound can be specified inside the angle brackets. If the same type parameter needs more than one upper bound, we need a separate where-clause:

fun <T> cloneWhenGreater(list: List<T>, threshold: T): List<T>
    where T : Comparable,
          T : Cloneable {
  return list.filter { it > threshold }.map { it.clone() }
}