I'm trying to understand the cake pattern.
I found this gist:
https://gist.github.com/2127745
But I don't understand this syntax:
// Explicit dependency on User Repository
self: UserRepositoryComponent =>
Can someone explain it please?
That's what's known as a self-type annotation. It means that you can assume that the object for the class has the declared type (in this case UserRepositoryComponent, or some subtype), and also (as a bonus) lets you refer to the "this" object at that level as "self", or whatever other name you specify. The self-type annotation is subtly powerful. It expresses a requirement for any implementation of the class (an earlier version of Scala expressed that syntactically as "requires UserRepositoryComponent") but doesn't actually imply a publicly visible type constraint (which would happen if you had said "extends UserRepositoryComponent"). The implementation requirement is enforced at any instantiation of the annotated class, but nowhere else. Self-type annotations are key to the "cake pattern", an encoding of program modules as Scala objects.
It's a self type annotation, explained for example here. If a self-type is given, it is taken as the type of this inside the trait. It lets Trait RealUserServiceComponent, via the self:UserRepositoryComment declaration say it can only be applied to classes that extend the UserRepositoryComment trait.
This is self-type annotation. It's a specification of what can extend your type. Like 'extends', but in the opopposite direction. Knowing that all your subtypes are some implementations of UserRepositoryComponent gives you some additional bonus. Namely you can call all of its methods as they were your own.
Related
What is the advantage of using an abstract class instead of a trait (apart from performance)? It seems like abstract classes can be replaced by traits in most cases.
I can think of two differences
Abstract classes can have constructor parameters as well as type parameters. Traits can have only type parameters. There was some discussion that in future even traits can have constructor parameters
Abstract classes are fully interoperable with Java. You can call them from Java code without any wrappers. Traits are fully interoperable only if they do not contain any implementation code
There's a section in Programming in Scala called "To trait, or not to trait?" which addresses this question. Since the 1st ed is available online, I'm hoping it's OK to quote the whole thing here. (Any serious Scala programmer should buy the book):
Whenever you implement a reusable collection of behavior, you will
have to decide whether you want to use a trait or an abstract class.
There is no firm rule, but this section contains a few guidelines to
consider.
If the behavior will not be reused, then make it a concrete class. It
is not reusable behavior after all.
If it might be reused in multiple, unrelated classes, make it a trait.
Only traits can be mixed into different parts of the class hierarchy.
If you want to inherit from it in Java code, use an abstract class.
Since traits with code do not have a close Java analog, it tends to be
awkward to inherit from a trait in a Java class. Inheriting from a
Scala class, meanwhile, is exactly like inheriting from a Java class.
As one exception, a Scala trait with only abstract members translates
directly to a Java interface, so you should feel free to define such
traits even if you expect Java code to inherit from it. See Chapter 29
for more information on working with Java and Scala together.
If you plan to distribute it in compiled form, and you expect outside
groups to write classes inheriting from it, you might lean towards
using an abstract class. The issue is that when a trait gains or loses
a member, any classes that inherit from it must be recompiled, even if
they have not changed. If outside clients will only call into the
behavior, instead of inheriting from it, then using a trait is fine.
If efficiency is very important, lean towards using a class. Most Java
runtimes make a virtual method invocation of a class member a faster
operation than an interface method invocation. Traits get compiled to
interfaces and therefore may pay a slight performance overhead.
However, you should make this choice only if you know that the trait
in question constitutes a performance bottleneck and have evidence
that using a class instead actually solves the problem.
If you still do not know, after considering the above, then start by
making it as a trait. You can always change it later, and in general
using a trait keeps more options open.
As #Mushtaq Ahmed mentioned, a trait cannot have any parameters passed to the primary constructor of a class.
Another difference is the treatment of super.
The other difference between classes and traits is that whereas in classes, super calls are statically bound, in traits, they are dynamically bound. If you write super.toString in a class, you know exactly which method implementation will be invoked. When you write the same thing in a trait, however, the method implementation to invoke for the super call is undefined when you define the trait.
See the rest of Chapter 12 for more details.
Edit 1 (2013):
There is a subtle difference in the way abstract classes behaves compared to traits. One of the linearization rules is that it preserves the inheritance hierarchy of the classes, which tends to push abstract classes later in the chain while traits can happily be mixed in. In certain circumstances, it's actually preferable to be in latter position of the class linearization, so abstract classes could be used for that. See constraining class linearization (mixin order) in Scala.
Edit 2 (2018):
As of Scala 2.12, trait's binary compatibility behavior has changed. Prior to 2.12, adding or removing a member to the trait required recompilation of all classes that inherit the trait, even if the classes have not changed. This is due to the way traits were encoded in JVM.
As of Scala 2.12, traits compile to Java interfaces, so the requirement has relaxed a bit. If the trait does any of the following, its subclasses still require recompilation:
defining fields (val or var, but a constant is ok – final val without result type)
calling super
initializer statements in the body
extending a class
relying on linearization to find implementations in the right supertrait
But if the trait does not, you can now update it without breaking binary compatibility.
For whatever it is worth, Odersky et al's Programming in Scala recommends that, when you doubt, you use traits. You can always change them into abstract classes later on if needed.
Other than the fact that you cannot directly extend multiple abstract classes, but you can mixin multiple traits into a class, it's worth mentioning that traits are stackable, since super calls in a trait are dynamically bound (it is referring a class or trait mixed before current one).
From Thomas's answer in Difference between Abstract Class and Trait:
trait A{
def a = 1
}
trait X extends A{
override def a = {
println("X")
super.a
}
}
trait Y extends A{
override def a = {
println("Y")
super.a
}
}
scala> val xy = new AnyRef with X with Y
xy: java.lang.Object with X with Y = $anon$1#6e9b6a
scala> xy.a
Y
X
res0: Int = 1
scala> val yx = new AnyRef with Y with X
yx: java.lang.Object with Y with X = $anon$1#188c838
scala> yx.a
X
Y
res1: Int = 1
When extending an abstract class, this shows that the subclass is of a similar kind. This is not neccessarily the case when using traits, I think.
In Programming Scala the authors say that abstract classes make a classical object oriented "is-a" relationship while traits are a scala-way of composition.
Abstract classes can contain behaviour - They can parameterized with constructor args (which traits can't) and represent a working entity. Traits instead just represent a single feature, an interface of one functionality.
A class can inherit from multiple traits but only one abstract class.
Abstract classes can have constructor parameters as well as type parameters. Traits can have only type parameters. For example, you can’t say trait t(i: Int) { }; the i parameter is illegal.
Abstract classes are fully interoperable with Java. You can call them from Java code without any wrappers. Traits are fully interoperable only if they do not contain any implementation code.
It's said in the documentation that
In native JS types, all concrete definitions must have = js.native as body. Any other body will be handled as if it were = js.native, and a warning will be emitted. (In Scala.js 1.0.0, this will become an error.)
And that's correct. However I found that I can omit body at all (thus making definition abstract) and there is no warning and generated js seems to be the same as with js.native body.
So my question is: what's the difference between abstract definitions and concrete definitions with js.native body?
The difference is that an abstract definition is abstract, and, well, a concrete definition (with = js.native) is concrete, from Scala's type system point of view.
But then what? From the use site of the class or trait, is doesn't make a difference. This is similar to normal Scala (or Java): when using a method, it doesn't matter whether it is abstract or not.
So the real difference is on the definition site. In theory, choosing abstract or concrete boils down to this criterium:
Does this method have an actual implementation in JavaScript code (not only a documented contract)? If yes, it should be concrete; if not, it should be abstract.
Practically and pragmatically, note that an abstract method can only appear in an abstract class or a trait, and must be implemented in a subclass/subtrait.
In terms of facade, in a native class, most methods should be concrete (if not all). That is because in JS, classes usually have concrete methods. In fact, abstract methods do not even exist in JS. The only reasonable case of defining an abstract method in a native class is if the "contract/documentation" of that class stipulates than a) it should be subclassed and b) subclasses should implement a particular method (not implemented in the superclass). This documented contract is as close as JS can get to abstract methods.
In JS traits, methods should usually be abstract (and the traits themselves be #ScalaJSDefined rather than #js.native). That is because traits/interfaces themselves do not even exist in JS. They only exist by their documented contract, which specifies what methods must/will be implemented by classes that satisfy this interface.
The only reasonable use case for concrete methods in (#js.native) JS traits is for DRYness. If several classes of a native API implement the same (large) set of methods, it can be reasonable to gather those methods in a native trait. In order not to have to repeat their definitions in all classes, they can be made concrete in the trait (if they were abstract, the classes would need to provided a concrete version to satisfy the contract). Note that such traits cannot be extended by non-native (#ScalaJSDefined) JS classes.
In the cases where you don't want to figure out the above "theoretical" criterium, use the following rule of thumb:
Is the method in a native JS class? If yes, it is almost certainly concrete.
Is it in a JS trait? If yes, it is almost certainly abstract (and the trait should be #ScalaJSDefined).
I read in Martin Odersky's book that trait extends a superclass AnyRef. So it's like a class for me. I know I cannot use default constructor in the trait which I usually use in the class
class B(s: String)
At the same time, I can see that it's impossible to do something like this
trait A {
def this(s: String) {
super()
}
}
Compiler says:
Error:(14, 7) 'this' expected but 'super' found.
super()
^
Why is that so?
Straight from the Scala Language Specification:
A trait is a class that is meant to be added to some other class as a mixin. Unlike normal classes, traits cannot have constructor parameters. Furthermore, no constructor arguments are passed to the superclass of the trait. This is not necessary as traits are initialized after the superclass is initialized.
What it comes down to is:
traits boil down to java interfaces, which have no parameters, which is nice for java-interop.
While it may be theoretically possible to allow traits to have
constructor parameters, it makes the language more complex than it
needs to be. There is almost certainly a way to accomplish the same functionality, without the need for trait constructor parameters.
I'm puzzled to choose a trait or class when writing scala code.
At first, I have a controller which with several traits:
class MyController extends Controller
with TransactionSupport
with JsonConverterSupport
with LoggerSupport
In these traits, I defined some methods and fields which can be used in MyController directly.
But my friend says: when you extends or with a trait, it should be a that trait.
Look at the MyController, it is a Controller, but it isn't a TransactionSupport, not a JsonConverterSupport, not a LoggerSupport, so it should not with them.
So the code becomes:
class MyController(tranSupport: TransactionSupport,
jsonConverter: JsonConverterSupport,
loggerSupport: LoggerSupport) extends Controller
But I don't feel good about this code, it just seems strange.
I see traits used heavily in scala code, when should I use it or use classes to inject?
I'll refer you to Interfaces should be Adjectives. Though some traits may play the part of a class (and, therefore, be nouns and respect the "is-a" relationship), when used as mixins they'll tend to play the part of interfaces.
As an "adjective", the trait will add a qualifying property to whatever they are extending. For example, they may be Comparable or Serializable.
It can be a bit hard to find an adjective to fit -- what adjective would you use for LoggerSupport? -- so don't feel overly constrained by that. Just be aware that it is completely wrong to thing of traits as necessarily an "is-a" relationship.
I would try to avoid using traits to replace "has-a" relationships, though.
My opinion is that it doesn't have to be it. Mixing-in is a different concept than inheritance. Even though syntactically it is the same, it doesn't mean the same. Typical use case for mixing-in is logging just like you wrote. It doesn't mean that if your service class mixes-in a Logging trait that it is a logger. It's just a yet another way how to compose functionality into working objects.
Odersky proposes that if you are not sure and you can, use traits because they are more flexible. You can change trait to class in the future if you need.
Sometime when I feel that mixing-in trait doesn't look good, I use module pattern like this:
trait JsonConverterModule {
protected def jsonConverter: JsonConverter
protected trait JsonConverter {
def convert(in: Json): Json
}
}
class MyController extends Controller with JsonConverterModule {
private doSmth = jsonConverter.convert(...)
}
MyController in this case looks more like a Controller, and all Json-related stuff is hidden from MyController 'client'
Your first example with traits is the "cake pattern" and your second example is "constructor injection". Both are perfectly valid ways to do dependency injection in Scala. The cake pattern is powerful, you can inject type members, the different traits can easily talk to each other (we don't have to create separate objects and pass them to each other object, often requiring setter injection rather than simple constructor injection), etc. However, the type has to be realized at compile-time, and a separate class must be realized for every combination of traits. Constructor injection lets you build your object at run-time and scales better for a large number of combinations.
What is the advantage of using an abstract class instead of a trait (apart from performance)? It seems like abstract classes can be replaced by traits in most cases.
I can think of two differences
Abstract classes can have constructor parameters as well as type parameters. Traits can have only type parameters. There was some discussion that in future even traits can have constructor parameters
Abstract classes are fully interoperable with Java. You can call them from Java code without any wrappers. Traits are fully interoperable only if they do not contain any implementation code
There's a section in Programming in Scala called "To trait, or not to trait?" which addresses this question. Since the 1st ed is available online, I'm hoping it's OK to quote the whole thing here. (Any serious Scala programmer should buy the book):
Whenever you implement a reusable collection of behavior, you will
have to decide whether you want to use a trait or an abstract class.
There is no firm rule, but this section contains a few guidelines to
consider.
If the behavior will not be reused, then make it a concrete class. It
is not reusable behavior after all.
If it might be reused in multiple, unrelated classes, make it a trait.
Only traits can be mixed into different parts of the class hierarchy.
If you want to inherit from it in Java code, use an abstract class.
Since traits with code do not have a close Java analog, it tends to be
awkward to inherit from a trait in a Java class. Inheriting from a
Scala class, meanwhile, is exactly like inheriting from a Java class.
As one exception, a Scala trait with only abstract members translates
directly to a Java interface, so you should feel free to define such
traits even if you expect Java code to inherit from it. See Chapter 29
for more information on working with Java and Scala together.
If you plan to distribute it in compiled form, and you expect outside
groups to write classes inheriting from it, you might lean towards
using an abstract class. The issue is that when a trait gains or loses
a member, any classes that inherit from it must be recompiled, even if
they have not changed. If outside clients will only call into the
behavior, instead of inheriting from it, then using a trait is fine.
If efficiency is very important, lean towards using a class. Most Java
runtimes make a virtual method invocation of a class member a faster
operation than an interface method invocation. Traits get compiled to
interfaces and therefore may pay a slight performance overhead.
However, you should make this choice only if you know that the trait
in question constitutes a performance bottleneck and have evidence
that using a class instead actually solves the problem.
If you still do not know, after considering the above, then start by
making it as a trait. You can always change it later, and in general
using a trait keeps more options open.
As #Mushtaq Ahmed mentioned, a trait cannot have any parameters passed to the primary constructor of a class.
Another difference is the treatment of super.
The other difference between classes and traits is that whereas in classes, super calls are statically bound, in traits, they are dynamically bound. If you write super.toString in a class, you know exactly which method implementation will be invoked. When you write the same thing in a trait, however, the method implementation to invoke for the super call is undefined when you define the trait.
See the rest of Chapter 12 for more details.
Edit 1 (2013):
There is a subtle difference in the way abstract classes behaves compared to traits. One of the linearization rules is that it preserves the inheritance hierarchy of the classes, which tends to push abstract classes later in the chain while traits can happily be mixed in. In certain circumstances, it's actually preferable to be in latter position of the class linearization, so abstract classes could be used for that. See constraining class linearization (mixin order) in Scala.
Edit 2 (2018):
As of Scala 2.12, trait's binary compatibility behavior has changed. Prior to 2.12, adding or removing a member to the trait required recompilation of all classes that inherit the trait, even if the classes have not changed. This is due to the way traits were encoded in JVM.
As of Scala 2.12, traits compile to Java interfaces, so the requirement has relaxed a bit. If the trait does any of the following, its subclasses still require recompilation:
defining fields (val or var, but a constant is ok – final val without result type)
calling super
initializer statements in the body
extending a class
relying on linearization to find implementations in the right supertrait
But if the trait does not, you can now update it without breaking binary compatibility.
For whatever it is worth, Odersky et al's Programming in Scala recommends that, when you doubt, you use traits. You can always change them into abstract classes later on if needed.
Other than the fact that you cannot directly extend multiple abstract classes, but you can mixin multiple traits into a class, it's worth mentioning that traits are stackable, since super calls in a trait are dynamically bound (it is referring a class or trait mixed before current one).
From Thomas's answer in Difference between Abstract Class and Trait:
trait A{
def a = 1
}
trait X extends A{
override def a = {
println("X")
super.a
}
}
trait Y extends A{
override def a = {
println("Y")
super.a
}
}
scala> val xy = new AnyRef with X with Y
xy: java.lang.Object with X with Y = $anon$1#6e9b6a
scala> xy.a
Y
X
res0: Int = 1
scala> val yx = new AnyRef with Y with X
yx: java.lang.Object with Y with X = $anon$1#188c838
scala> yx.a
X
Y
res1: Int = 1
When extending an abstract class, this shows that the subclass is of a similar kind. This is not neccessarily the case when using traits, I think.
In Programming Scala the authors say that abstract classes make a classical object oriented "is-a" relationship while traits are a scala-way of composition.
Abstract classes can contain behaviour - They can parameterized with constructor args (which traits can't) and represent a working entity. Traits instead just represent a single feature, an interface of one functionality.
A class can inherit from multiple traits but only one abstract class.
Abstract classes can have constructor parameters as well as type parameters. Traits can have only type parameters. For example, you can’t say trait t(i: Int) { }; the i parameter is illegal.
Abstract classes are fully interoperable with Java. You can call them from Java code without any wrappers. Traits are fully interoperable only if they do not contain any implementation code.