Scala Trait 的线性化

Scala Trait调用super方法的线性化是一个难理解的内容，下面让我通过一段代码来推演线性化过程。

源代码：

package c5soft

class SuperBase {

def show() {

println("Here step to class SuperBase")

println("I'm " + toString + "\n")

}

}

trait TraitBase extends SuperBase {

override def toString = "trait TraitBase"

override def show() {

println("Here step to trait TraitBase")

super.show()

}

}

trait A extends TraitBase {

override def toString = "trait A"

override def show() {

println("Here step to trait A")

super.show()

}

}

trait AChild extends A {

override def toString = "trait AChild"

override def show() {

println("Here step to trait AChild")

super.show()

}

}

trait B extends TraitBase {

override def toString = "trait B"

override def show() {

println("Here step to trait B")

super.show()

}

}

trait BChild extends B {

override def toString = "trait BChild"

override def show() {

println("Here step to trait BChild")

super.show()

}

}

class C extends SuperBase {

override def toString = "class C"

override def show() {

println("Here step to class C")

super.show()

}

}

object Main extends App {

println("----------C.show-----")

new C show()

println("----------CA.show-----")

new C with A show()

println("----------CB.show-----")

new C with B show()

println("---------CAB.show----")

new C with A with B show()

println("---------CaB.show----")

new C with AChild with B show()

println("---------CBA.show----")

new C with B with A show()

println("---------CBa.show----")

new C with B with AChild show()

println("---------Cab.show----")

new C with AChild with BChild show()

println("---------Ca.show----")

new C with AChild show()

}

输出结果：

----------C.show-----

Here step to class C

Here step to class SuperBase

I'm class C

----------CA.show-----

Here step to trait A

Here step to trait TraitBase

Here step to class C

Here step to class SuperBase

I'm trait A

----------CB.show-----

Here step to trait B

Here step to trait TraitBase

Here step to class C

Here step to class SuperBase

I'm trait B

---------CAB.show----

Here step to trait B

Here step to trait A

Here step to trait TraitBase

Here step to class C

Here step to class SuperBase

I'm trait B

---------CaB.show----

Here step to trait B

Here step to trait AChild

Here step to trait A

Here step to trait TraitBase

Here step to class C

Here step to class SuperBase

I'm trait B

---------CBA.show----

Here step to trait A

Here step to trait B

Here step to trait TraitBase

Here step to class C

Here step to class SuperBase

I'm trait A

---------CBa.show----

Here step to trait AChild

Here step to trait A

Here step to trait B

Here step to trait TraitBase

Here step to class C

Here step to class SuperBase

I'm trait AChild

---------Cab.show----

Here step to trait BChild

Here step to trait B

Here step to trait AChild

Here step to trait A

Here step to trait TraitBase

Here step to class C

Here step to class SuperBase

I'm trait BChild

---------Ca.show----

Here step to trait AChild

Here step to trait A

Here step to trait TraitBase

Here step to class C

Here step to class SuperBase

I'm trait AChild

仔细分析这段代码与输出结果，你就能明了Trait的super调用是如何线性化的。

已CBa.show为例:

println("---------CBa.show----")       //小写的a代表AChild

new C with B with AChild show()  //这里等同于(new C with B with AChild).show()

调用过程从右到左并依次调用其基类的过程为：

Here step to trait AChild

Here step to trait A

Here step to trait TraitBase

Here step to class SuperBase

Here step to trait B

Here step to trait TraitBase

Here step to class SuperBase

Here step to class C

Here step to class SuperBase

按照【1.基类调用只出现一次；2.后面出现的基类调用在前面删除】原则处理如下：

Here step to trait AChild

Here step to trait A

Here step to trait TraitBase

Here step to class SuperBase

Here step to trait B

Here step to trait TraitBase

Here step to class SuperBase

Here step to class C

Here step to class SuperBase

最终得到:

Here step to trait AChild

Here step to trait A

Here step to trait B

Here step to trait TraitBase

Here step to class C

Here step to class SuperBase