Scala中的语言特性是如何实现的(3) -- trait

我在Coursera上跟了一门叫做Functional Programming Principles in Scala的课程,是由Scala的作者Martin Odersky讲授的。其中第三周的作业中使用到了Scala的trait这个语言特性。

我以前熟知的语言都没有类似的特性(Ruby的mixin和Scala的trait很像,但是Ruby我不熟),所以这周的博客就分析一下这个语言特性是如何实现的。

trait

在讲trait的实现机制之前,先看一个使用trait的例子。
假设我们有以下几个类:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
abstract class Plant {
def photosynthesis = println("Oh, the sunlight!")
}

class Rose extends Plant {
def smell = println("Good!")

def makePeopleHappy = println("People like me")
}

class Ruderal extends Plant {
def grow = println("I take up all the space!")
}

abstract class Animal {
def move = println("I can move!")
}

class Dog extends Animal {
def bark = println("Woof!")

def makePeopleHappy = println("People like me")
}

class Snake extends Animal {
def bite = println("I am poisonous!")
}

植物家族有玫瑰和杂草。

动物家族有狗和毒蛇。

仔细观察可以发现,玫瑰和狗有一个共同的行为,它们都可以取悦人类,这个行为是用完全一样的代码实现的。

如何把Rose和Dog中的重复代码消除掉呢?有一种潜在的解决方案:
把makePeopleHappy提取到一个类中去,让植物和动物都继承自它。

这么做虽然消除了重复代码但有两个明显的缺点:

  1. 植物和动物继承自同一个类,不太合理
  2. 杂草和毒蛇也具有了取悦于人的能力,也不太合理

这时我们就可以使用trait,它没有上面提到的两个缺点。

1
2
3
4
5
6
7
8
9
10
11
trait PeoplePleaser {
def makePeopleHappy = println("People like me")
}

class Rose extends Plant with PeoplePleaser {
def smell = println("Good!")
}

class Dog extends Animal with PeoplePleaser {
def bark = println("Woof!")
}

我们定义一个trait,把makePeopleHappy置于其中,让Rose和Dog都with这个trait。然后就可以写这样的代码来调用它们了:

1
2
new Rose().makePeopleHappy
new Dog().makePeopleHappy

这样我们就解决了重复代码的问题,而且没有触及已存在的继承关系。

现在看看trait的实现机制吧,我们开始反编译!

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
public abstract interface PeoplePleaser
{
public abstract void makePeopleHappy();
}

public abstract class PeoplePleaser$class
{
public static void makePeopleHappy(PeoplePleaser $this)
{
Predef..MODULE$.println("People like me");
}

public static void $init$(PeoplePleaser $this)
{
}
}

public class Rose extends Plant
implements PeoplePleaser
{
public void makePeopleHappy()
{
PeoplePleaser$class.makePeopleHappy(this);
}

public void smell() { Predef..MODULE$.println("Good!"); }

public Rose()
{
PeoplePleaser.class.$init$(this);
}
}

public class Dog extends Animal
implements PeoplePleaser
{
public void makePeopleHappy()
{
PeoplePleaser$class.makePeopleHappy(this);
}

public void bark() { Predef..MODULE$.println("Woof!"); }

public Dog()
{
PeoplePleaser.class.$init$(this);
}
}

真相大白了,PeoplePleaser被编译成了一个接口加一个抽象类。Rose和Dog实现这个接口,并通过调用抽象类中的静态方法来实现了makePeopleHappy。

很有趣的一点是Rose和Dog在调用静态方法时都把this传了进去,为什么呢?我们把原来的代码改成这样来看:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
trait PeoplePleaser {
val moreMessage = ""

def makePeopleHappy = println("People like me. " + moreMessage)
}

class Rose extends Plant with PeoplePleaser {
override val moreMessage = "Because I smell nice."

def smell = println("Good!")
}

class Dog extends Animal with PeoplePleaser {
override val moreMessage = "Because I fetch balls."

def bark = println("Woof!")
}

我们给makePeopleHappy加上一段额外的信息。
现在再次反编译。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
public abstract interface PeoplePleaser
{
public abstract void objsets$PeoplePleaser$_setter_$moreMessage_$eq(String paramString);

public abstract String moreMessage();

public abstract void makePeopleHappy();
}

public abstract class PeoplePleaser$class
{
public static void makePeopleHappy(PeoplePleaser $this)
{
Predef..MODULE$.println(new StringBuilder()
.append("People like me. ")
.append($this.moreMessage()).toString());
}

public static void $init$(PeoplePleaser $this)
{
$this.objsets$PeoplePleaser$_setter_$moreMessage_$eq("");
}
}

public class Rose extends Plant
implements PeoplePleaser
{
private final String moreMessage;

public void objsets$PeoplePleaser$_setter_$moreMessage_$eq(String x$1)
{
}

public void makePeopleHappy()
{
PeoplePleaser$class.makePeopleHappy(this);
}

public String moreMessage() { return this.moreMessage; }

public void smell() {
Predef..MODULE$.println("Good!");
}

public Rose()
{
PeoplePleaser.class.$init$(this);
this.moreMessage = "Because I smell nice.";
}
}

public class Dog extends Animal
implements PeoplePleaser
{
private final String moreMessage;

public void objsets$PeoplePleaser$_setter_$moreMessage_$eq(String x$1)
{
}

public void makePeopleHappy()
{
PeoplePleaser$class.makePeopleHappy(this);
}

public String moreMessage() { return this.moreMessage; }

public void bark() {
Predef..MODULE$.println("Woof!");
}

public Dog()
{
PeoplePleaser.class.$init$(this);
this.moreMessage = "Because I fetch balls.";
}
}

现在就清楚了,抽象类中的静态方法可能会依赖于各个实例不同的状态,所以需要把this传递进去。
这样我们才能够给makePeopleHappy加上一段额外的信息。