我们可以在什么地方创建对象?
-
全局变量区
Person p;
-
栈
void Max() { Person p; }
-
堆
new
和delete
//在堆中创建对象: Person* p = new Person(); //释放对象占用的内存 delete p;
在堆中创建对象: new``delete
在C语言中我们使用malloc
申请堆空间
使用完毕后使用free
释放空间
C++:
class Person
{
private:
int x;
int y;
public:
Person()
{
printf("Person()执行了! \n");
}
Person(int x,int y)
{
printf("Person(int x,int y)执行了! \n");
this->x=x;
this->y=y;
}
~Person()
{
printf("~Person() 执行了! \n");
}
};
Person* p =new Person();
Person* p1=new Person(1,2);
delete p;
delete p1;
//反汇编
33: Person* p =new Person();
00401070 push 8
00401072 call operator new (004013c0)
00401077 add esp,4
0040107A mov dword ptr [ebp-1Ch],eax
0040107D mov dword ptr [ebp-4],0
00401084 cmp dword ptr [ebp-1Ch],0
00401088 je main+57h (00401097)
0040108A mov ecx,dword ptr [ebp-1Ch]
0040108D call @ILT+20(Person::Person) (00401019)
00401092 mov dword ptr [ebp-38h],eax
00401095 jmp main+5Eh (0040109e)
00401097 mov dword ptr [ebp-38h],0
0040109E mov eax,dword ptr [ebp-38h]
004010A1 mov dword ptr [ebp-18h],eax
004010A4 mov dword ptr [ebp-4],0FFFFFFFFh
004010AB mov ecx,dword ptr [ebp-18h]
004010AE mov dword ptr [ebp-10h],ecx
34: Person* p1=new Person(1,2);
004010B1 push 8
004010B3 call operator new (004013c0)
004010B8 add esp,4
004010BB mov dword ptr [ebp-24h],eax
004010BE mov dword ptr [ebp-4],1
004010C5 cmp dword ptr [ebp-24h],0
004010C9 je main+9Ch (004010dc)
004010CB push 2
004010CD push 1
004010CF mov ecx,dword ptr [ebp-24h]
004010D2 call @ILT+0(Person::Person) (00401005)
004010D7 mov dword ptr [ebp-3Ch],eax
004010DA jmp main+0A3h (004010e3)
004010DC mov dword ptr [ebp-3Ch],0
004010E3 mov edx,dword ptr [ebp-3Ch]
004010E6 mov dword ptr [ebp-20h],edx
004010E9 mov dword ptr [ebp-4],0FFFFFFFFh
004010F0 mov eax,dword ptr [ebp-20h]
004010F3 mov dword ptr [ebp-14h],eax
35: delete p;
004010F6 mov ecx,dword ptr [ebp-10h]
004010F9 mov dword ptr [ebp-2Ch],ecx
004010FC mov edx,dword ptr [ebp-2Ch]
004010FF mov dword ptr [ebp-28h],edx
00401102 cmp dword ptr [ebp-28h],0
00401106 je main+0D7h (00401117)
00401108 push 1
0040110A mov ecx,dword ptr [ebp-28h]
0040110D call @ILT+15(Person::`scalar deleting destructor") (00401014)
00401112 mov dword ptr [ebp-40h],eax
00401115 jmp main+0DEh (0040111e)
00401117 mov dword ptr [ebp-40h],0
36: delete p1;
0040111E mov eax,dword ptr [ebp-14h]
00401121 mov dword ptr [ebp-34h],eax
00401124 mov ecx,dword ptr [ebp-34h]
00401127 mov dword ptr [ebp-30h],ecx
0040112A cmp dword ptr [ebp-30h],0
0040112E je main+0FFh (0040113f)
00401130 push 1
00401132 mov ecx,dword ptr [ebp-30h]
00401135 call @ILT+15(Person::`scalar deleting destructor") (00401014)
0040113A mov dword ptr [ebp-44h],eax
0040113D jmp main+106h (00401146)
0040113F mov dword ptr [ebp-44h],0
new
关键字到底做了什么?
new``delete
的本质:
-
分析
malloc
函数的执行流程
char* p = (char*)malloc(123);
//逐层调用 00401072 call malloc (004013f0) 00401404 call _nh_malloc_dbg (00401470) 00401484 call _heap_alloc_dbg (004014e0) // 00401559 call dword ptr [__pfnAllocHook (00427cb0)] 0040167D call _heap_alloc_base (00404380) 0040443C call dword ptr [__imp__HeapAlloc@12 (0042d190)]
-
分析
new
的执行流程
Person p = new Person();
36: Person* p =new Person(); 0040107D push 8 0040107F call operator new (004013d0) //new //逐层调用 004013DA call _nh_malloc (00401450) 00401461 call _nh_malloc_dbg (00401470) 00401484 call _heap_alloc_dbg (004014e0) // 00401559 call dword ptr [__pfnAllocHook (00427cb0)] 0040167D call _heap_alloc_base (00404380) 0040443C call dword ptr [__imp__HeapAlloc@12 (0042d190)] // 00401084 add esp,4 00401087 mov dword ptr [ebp-20h],eax 0040108A mov dword ptr [ebp-4],0 00401091 cmp dword ptr [ebp-20h],0 00401095 je main+64h (004010a4) 00401097 mov ecx,dword ptr [ebp-20h] 0040109A call @ILT+20(Person::Person) (00401019) 0040109F mov dword ptr [ebp-3Ch],eax 004010A2 jmp main+6Bh (004010ab) 004010A4 mov dword ptr [ebp-3Ch],0 004010AB mov eax,dword ptr [ebp-3Ch] 004010AE mov dword ptr [ebp-1Ch],eax 004010B1 mov dword ptr [ebp-4],0FFFFFFFFh 004010B8 mov ecx,dword ptr [ebp-1Ch] 004010BB mov dword ptr [ebp-14h],ecx
总结:
new = malloc + 构造函数
动动手:
分析 delete的执行流程 和free对比
数组堆空间申请 new[]/delete[]
分别用C和C++方式在堆中申请Int数组
int* p = (int*)malloc(sizeof(int)*10); free(p);
int* p = new int[10]; delete[]p;
分别用C和C++方式在堆栈中申请Class类型数组
int* p = (Person*)malloc(sizeof(Person)*10); free(p);
Person* p = new Person[10]; delete[] p;
delete和delete[]有什么区别?
如果对象数组 只使用delete p
(一个delete)我们发现 只有一个析构函数执行
要把10个对象所占用空间全部释放,并且每个都要执行析构函数的话 必须使用 delete[] p
引用类型
引用类型是C++中的特性
引用类型就是变量的别名
-
基本类型
int x= 1; int& p =x; //起个别名叫p p = 2; //p就是x printf("%d \n",x);
-
类
Person p; Person& px =p //起个别名叫px px.x = 10; //px 就是p printf("%d \n",p.x);
-
指针
int****** i = (int******)1; int******& r =i; //起个别名叫r r = (int******)2; //r就是I printf("%d \n",r);
-
数组
int arr[] = {1,2,3}; int (&p)[3] = arr; //起个别名叫 p p[0] =4; //p 就是arr printf("%d \n",arr[0]);
引用类型就是给变量起的别名
引用类型在定义时必须赋初始值
引用类型的本质
int x= 1;
int& p =x; //起个别名叫p
p = 2; //p就是x
printf("%d \n",x);
43: int x= 1;
00401058 mov dword ptr [ebp-4],1
44: int& p =x; //起个别名叫p
0040105F lea eax,[ebp-4]
00401062 mov dword ptr [ebp-8],eax
45: p = 2; //p就是x
00401065 mov ecx,dword ptr [ebp-8]
00401068 mov dword ptr [ecx],2
46: printf("%d \n",x);
0040106E mov edx,dword ptr [ebp-4]
00401071 push edx
00401072 push offset string "%d \n" (0042501c)
00401077 call printf (00403880)
0040107C add esp,8
//测试 吧上述代码中的引用类型改为指针。
int x= 1;
int* p =&x; //起个别名叫p
*p = 2; //p就是x
printf("%d \n",x);
//结果发现生成的反汇编结果一模一样
//这里我们暂时得出结论,,引用类型就是指针。或者说在底层实现上 就是指针
引用类型和指针的区别
int x = 1;
//必须初始化
int* p = &x;
int& ref = x;
//运算
p++;
ref++;
//赋值
p = (int*)1;
ref = 100;
//反汇编
49: int x = 1;
00401028 mov dword ptr [ebp-4],1
50: //必须初始化
51: int* p = &x;
0040102F lea eax,[ebp-4]
00401032 mov dword ptr [ebp-8],eax
52: int& ref = x;
00401035 lea ecx,[ebp-4]
00401038 mov dword ptr [ebp-0Ch],ecx
53:
54: //运算
55: p++;
0040103B mov edx,dword ptr [ebp-8]
0040103E add edx,4
00401041 mov dword ptr [ebp-8],edx
56: ref++;
00401044 mov eax,dword ptr [ebp-0Ch]
00401047 mov ecx,dword ptr [eax]
00401049 add ecx,1
0040104C mov edx,dword ptr [ebp-0Ch]
0040104F mov dword ptr [edx],ecx
57:
58: //赋值
59: p = (int*)1;
00401051 mov dword ptr [ebp-8],1
60: ref = 100;
00401058 mov eax,dword ptr [ebp-0Ch]
0040105B mov dword ptr [eax],64h
class Base
{
public:
int x;
int y;
Base(int x,int y)
{
this->x = x;
this->y =y;
}
}
Base b(1,2);
//必须初始化
Base* p = &b;
Base& ref = b;
//运算
p++;
//ref++;
//赋值
p = (Base*)1;
//ref = 100;
总结:
- 被引用必须赋初始值,且只能指向一个变量,“从一而终”。
- 对引用复制,是对其指向的变量赋值,而不是修改引用本身的值
- 对引用做运算,就是对其指向的变量做运算,而不是对引用本身做运算。
- 引用类型就是一个“弱化了的指针”。
个人见解,其实我觉得引用类型就像是一个
*p
也就是取值了的指针~~~
引用在函数参数传递中的作用(基本类型)
void Plus(int& i)
{
i++;
return;
}
int main(int argc,char* argv[])
{
int i = 10;
Plus(i);
printf("%d \n",i);
return 0;
}
102: i++;
00401038 mov eax,dword ptr [ebp+8]
0040103B mov ecx,dword ptr [eax]
0040103D add ecx,1
00401040 mov edx,dword ptr [ebp+8]
00401043 mov dword ptr [edx],ecx
//传入的虽然是指针,但是对参数操作就是对对应参数操作,而不是指针本身的操作
108: int i = 10;
00401078 mov dword ptr [ebp-4],0Ah
109: Plus(i);
0040107F lea eax,[ebp-4]
00401082 push eax
00401083 call @ILT+0(Plus) (00401005)
00401088 add esp,4
110: printf("%d \n",i);
0040108B mov ecx,dword ptr [ebp-4]
0040108E push ecx
0040108F push offset string "%d \n" (0042201c)
00401094 call printf (004010d0)
00401099 add esp,8
//当函数的参数为引用类型的时候
//传入的参数为参数的地址
引用在函数参数传递中的作用(构造类型)
struct Base
{
int x;
int y;
Base(int x,int y)
{
this->x = x;
this->y = y;
}
}
void PrintByRef(Base& refb,Base* pb)
{
//通过指针读取
printf("%d,%d \n",pb->x,pb->y);
//通过引用获取
printf("%d %d \n",refb.x,refb.y);
//指针可以重新赋值,可以做运算
//引用不可以
//refb = (Base&)1;
//refb++;
}
//反汇编
121: //通过指针读取
122: printf("%d,%d \n",pb->x,pb->y);
00401088 mov eax,dword ptr [ebp+0Ch]
0040108B mov ecx,dword ptr [eax+4]
0040108E push ecx
0040108F mov edx,dword ptr [ebp+0Ch]
00401092 mov eax,dword ptr [edx]
00401094 push eax
00401095 push offset string "%d,%d \n" (00422028)
0040109A call printf (004011b0)
0040109F add esp,0Ch
123: //通过引用获取
124: printf("%d %d \n",refb.x,refb.y);
004010A2 mov ecx,dword ptr [ebp+8]
004010A5 mov edx,dword ptr [ecx+4]
004010A8 push edx
004010A9 mov eax,dword ptr [ebp+8]
004010AC mov ecx,dword ptr [eax]
004010AE push ecx
004010AF push offset string "%d %d \n" (0042201c)
004010B4 call printf (004011b0)
004010B9 add esp,0Ch
125: //指针可以重新赋值,可以做运算
126: //引用不可以
127: //refb = (Base&)1;
128: //refb++;
133: Base b(1,2);
00401108 push 2
0040110A push 1
0040110C lea ecx,[ebp-8]
0040110F call @ILT+5(Base::Base) (0040100a)
134: Base* p =&b;
00401114 lea eax,[ebp-8]
00401117 mov dword ptr [ebp-0Ch],eax
135: Base& ref=b;
0040111A lea ecx,[ebp-8]
0040111D mov dword ptr [ebp-10h],ecx
136:
137: PrintByRef(ref,p);
00401120 mov edx,dword ptr [ebp-0Ch]
00401123 push edx
00401124 mov eax,dword ptr [ebp-10h]
00401127 push eax
00401128 call @ILT+15(PrintByRef) (00401014)
0040112D add esp,8
给狗起个人的名字?
引用是变量的别名,如:
int x = 10;
int& r = x; //int类型的别名就应该是 Int&
Base b(1,2);
Base&r = b; //Base类型的别名 就应该是 Base&
Base& r = (Base&)x; //虽然可以变异 但是意义不大
常引用
class Base
{
public:
int x;
};
void Print(const Base& ref)//常量参数
{
//ref = 100; //不论是不是const不能修改
//ref.x = 200; //不是const能修改指向的内容 是const 不能修改
printf("%d \n",ref.x);
}
int main(int argc,char* argv[])
{
Base b;
b.x = 100;
Print(b);
return 0;
}