关于C++入门(5)练习13.13个问题

Question

我的代码如下：

#include <iostream>
#include <vector>
#include <string>

using std::cout;
using std::cin;
using std::endl;
using std::vector;
using std::string;

struct X
{
    X() { cout << "X()" << endl; }
    X( const string &s ):item( s ) { cout << "X( initialize by str(" << s << ") )" << endl; }
    X( const& X ) { cout << "X( const X& )" << endl; }
    ~X() { cout << "X(" << (*this).item << ") was destroy" << endl; }

    string item;
};

int main()
{

    X a( "aa" );
    X b( "bb" );
    X c( "cc" );

    vector<X> vec;

    vec.push_back( a );
    vec.push_back( b );
    vec.push_back( c );

    vec[0].item = "vec_aa";
    vec[1].item = "vec_bb";
    vec[2].item = "vec_cc";

}

当我运行这个程序时，我得到了这个输出：

X(由str(aa)初始化)

X(由str(bb)初始化)

X(由str(cc)初始化)

X(aa)被摧毁

X(aa)被摧毁

X(bb)被摧毁

X(vec_aa)被摧毁

X(vec_bb)被摧毁

X(vec_cc)被摧毁

X(cc)被摧毁

X(bb)被摧毁

X(aa)被摧毁

所以，我的问题是:它为什么输出

X(aa) was destroy  
X(aa) was destroy  
X(bb) was destroy  

我预期输出是

X(aa) was destroy  
X(bb) was destroy  
X(cc) was destroy 

Answer 1

std::vector驻留在连续内存中--如果还不够大，就需要重新调整大小，这意味着在新内存中分配新内存、移动(如果可用)或复制构造新内存中的对象，并在删除之前销毁旧内存中的对象。

显然，这样的重新分配会发生两次，这就是您意外的输出的来源：

X(aa) was destroy // first re-allocation: copy of aa resides in vector with
                  // capacity of size 1, but needs to be 2

X(aa) was destroy // second re-allocation: copies of aa and bb reside in
X(bb) was destroy // vector with capacity of 2, but needs to be 3
                  // (will, most likely, actually get 4)

您可以避免这一点，如果您预先reserve足够的内存-这是您应该做的，如果您事先知道(即使只是粗略地)有多少对象，您将放置到向量中，因为重新分配是昂贵的。