函数调用中的一元指针增量与函数调用之前/之后的增量

发布于 2024-08-16 01:27:03 字数 6202 浏览 11 评论 0原文

我试图理解一段代码,这是导致混乱的片段:

typedef map<int, Person, less<int> > people_map;
people_map people;
.
.
.
cout << "Erasing people of age 100" << endl;

    for (people_map::iterator j = people.begin(); j != people.end();) {
        if (j->second.GetAge() == 100)
        {

            people.erase(j++); // iterator is advanced before the erase occurs

        }
        else
            ++j; // advance the iterator
    } // end of erase loop

混乱是:如果我想在函数调用后增加 j ,它会导致分段错误。我无法理解为什么:

我将其更改为这样的内容:

 if (j->second.GetAge() == 100)
        {
            temp = j++;
            j--;
            people.erase(j); // iterator is advanced before the erase occurs
            j=temp;

        }

导致分段错误。

或像这样:

 if (j->second.GetAge() == 100)
        {
            people.erase(j); // iterator is advanced before the erase occurs
            j++;   
        }

导致分段错误。

这是完整的程序清单:

// disable warnings about long names
#ifdef WIN32
#pragma warning( disable : 4786)
#endif

#include <string>
#include <map>
#include <algorithm>
#include <iostream>
#include <sstream>
#include <iterator>
#include <functional>

using namespace std;

class Person {
    // private members
    string m_sName;
    string m_sEmail;
    int m_iAge;

public:

    // constructor

    Person(const string sName,
            const string sEmail,
            const int iAge) :
    m_sName(sName), m_sEmail(sEmail), m_iAge(iAge) {
    };

    // default constructor

    Person() : m_iAge(0) {
    };

    // copy constructor

    Person(const Person & p) :
    m_sName(p.m_sName), m_sEmail(p.m_sEmail), m_iAge(p.m_iAge) {
    };

    // operator =

    Person & operator=(const Person & rhs) {
        // don't assign to self
        if (this == &rhs)
            return *this;

        m_sName = rhs.m_sName;
        m_sEmail = rhs.m_sEmail;
        m_iAge = rhs.m_iAge;
        return *this;
    };

    // access private members

    string GetName() const {
        return m_sName;
    };

    string GetEmail() const {
        return m_sEmail;
    };

    int GetAge() const {
        return m_iAge;
    };

}; // end of class Person

// function object to print one person

class fPrint {
    ostream & m_os;

public:

    // constructor - remember which stream to use

    fPrint(ostream & os) : m_os(os) {
    };

    // person object arrives as a pair of key,object

    void operator() (const pair <const int, const Person> & item) const {
        m_os << "# " << item.first << " - name: "
                << item.second.GetName()
                << " - " << item.second.GetEmail()
                << ", age " << item.second.GetAge()
                << endl;
    };

}; // end of class fPrint

// declare type for storing people (numeric key, person object)
typedef map<int, Person, less<int> > people_map;

int main(void) {
    // make a map of people
    people_map people;

    // add items to list
    people [1234] = Person("Nick", "[email protected]", 15);
    people [4422] = Person("Fred", "[email protected]", 100);
    people [88] = Person("John", "[email protected]", 35);
    // insert a different way ...
    people.insert(make_pair(42, Person("Abigail", "[email protected]", 22)));

    // best to declare this on its own line :)
    fPrint fo(cout); // instance of function output object

    // print everyone (calls a function object to print)
    cout << "Printing all using fPrint ..." << endl;
    for_each(people.begin(), people.end(), fo);

    // find someone by key
    cout << "Finding person 4422 ..." << endl;

    people_map::const_iterator i = people.find(4422);

    if (i == people.end())
        cout << "Not found." << endl;
    else {
        fo(*i); // dereference and print

        // another way of printing -

        // key itself is the "first" part of the map pair ...
        cout << "Found key = " << i->first << endl;

        // person object is the "second" part of the map pair...

        cout << "Found name = " << i->second.GetName() << endl;
    }

    // Note, this will not work:
    //   fPrint (cout) (*i);

    // However this will:
    //   0, fPrint (cout) (*i);

    // However I think the extra zero is a bit obscure. :)

    // An alternative way of finding someone.
    // Note - this will add them if they are not there.
    // Since this is a reference changing it will change the person in the
    // map. Leave off the & to get a copy of the person.

    Person & p = people [1234];

    cout << "Person 1234 has name " << p.GetName() << endl;

    // Example of erasing an element correctly ...
    // If we did the j++ as part of the for loop we would end up
    // adding 1 to an iterator that pointed to an element that was
    // removed which would lead to a crash. See Josuttis p 205.

    cout << "Erasing people of age 100" << endl;

    for (people_map::iterator j = people.begin(); j != people.end();) {
        if (j->second.GetAge() == 100)
        {

            people.erase(j++); // iterator is advanced before the erase occurs

        }
        else
            ++j; // advance the iterator
    } // end of erase loop


    // now display who is left
    cout << "Printing people left after erase ..." << endl;
    for_each(people.begin(), people.end(), fo);

    return 0;
} // end of main

I am trying to understand a code, here is fragment which is causing confusion:

typedef map<int, Person, less<int> > people_map;
people_map people;
.
.
.
cout << "Erasing people of age 100" << endl;

    for (people_map::iterator j = people.begin(); j != people.end();) {
        if (j->second.GetAge() == 100)
        {

            people.erase(j++); // iterator is advanced before the erase occurs

        }
        else
            ++j; // advance the iterator
    } // end of erase loop

the confusion is: if i want to increment j after the function call it causes segmentation fault. I am not able to understand why:

I change it to something like this:

 if (j->second.GetAge() == 100)
        {
            temp = j++;
            j--;
            people.erase(j); // iterator is advanced before the erase occurs
            j=temp;

        }

causes segmentation fault.

or like this:

 if (j->second.GetAge() == 100)
        {
            people.erase(j); // iterator is advanced before the erase occurs
            j++;   
        }

causes segmentation fault.

here is the complete program listing:

// disable warnings about long names
#ifdef WIN32
#pragma warning( disable : 4786)
#endif

#include <string>
#include <map>
#include <algorithm>
#include <iostream>
#include <sstream>
#include <iterator>
#include <functional>

using namespace std;

class Person {
    // private members
    string m_sName;
    string m_sEmail;
    int m_iAge;

public:

    // constructor

    Person(const string sName,
            const string sEmail,
            const int iAge) :
    m_sName(sName), m_sEmail(sEmail), m_iAge(iAge) {
    };

    // default constructor

    Person() : m_iAge(0) {
    };

    // copy constructor

    Person(const Person & p) :
    m_sName(p.m_sName), m_sEmail(p.m_sEmail), m_iAge(p.m_iAge) {
    };

    // operator =

    Person & operator=(const Person & rhs) {
        // don't assign to self
        if (this == &rhs)
            return *this;

        m_sName = rhs.m_sName;
        m_sEmail = rhs.m_sEmail;
        m_iAge = rhs.m_iAge;
        return *this;
    };

    // access private members

    string GetName() const {
        return m_sName;
    };

    string GetEmail() const {
        return m_sEmail;
    };

    int GetAge() const {
        return m_iAge;
    };

}; // end of class Person

// function object to print one person

class fPrint {
    ostream & m_os;

public:

    // constructor - remember which stream to use

    fPrint(ostream & os) : m_os(os) {
    };

    // person object arrives as a pair of key,object

    void operator() (const pair <const int, const Person> & item) const {
        m_os << "# " << item.first << " - name: "
                << item.second.GetName()
                << " - " << item.second.GetEmail()
                << ", age " << item.second.GetAge()
                << endl;
    };

}; // end of class fPrint

// declare type for storing people (numeric key, person object)
typedef map<int, Person, less<int> > people_map;

int main(void) {
    // make a map of people
    people_map people;

    // add items to list
    people [1234] = Person("Nick", "[email protected]", 15);
    people [4422] = Person("Fred", "[email protected]", 100);
    people [88] = Person("John", "[email protected]", 35);
    // insert a different way ...
    people.insert(make_pair(42, Person("Abigail", "[email protected]", 22)));

    // best to declare this on its own line :)
    fPrint fo(cout); // instance of function output object

    // print everyone (calls a function object to print)
    cout << "Printing all using fPrint ..." << endl;
    for_each(people.begin(), people.end(), fo);

    // find someone by key
    cout << "Finding person 4422 ..." << endl;

    people_map::const_iterator i = people.find(4422);

    if (i == people.end())
        cout << "Not found." << endl;
    else {
        fo(*i); // dereference and print

        // another way of printing -

        // key itself is the "first" part of the map pair ...
        cout << "Found key = " << i->first << endl;

        // person object is the "second" part of the map pair...

        cout << "Found name = " << i->second.GetName() << endl;
    }

    // Note, this will not work:
    //   fPrint (cout) (*i);

    // However this will:
    //   0, fPrint (cout) (*i);

    // However I think the extra zero is a bit obscure. :)

    // An alternative way of finding someone.
    // Note - this will add them if they are not there.
    // Since this is a reference changing it will change the person in the
    // map. Leave off the & to get a copy of the person.

    Person & p = people [1234];

    cout << "Person 1234 has name " << p.GetName() << endl;

    // Example of erasing an element correctly ...
    // If we did the j++ as part of the for loop we would end up
    // adding 1 to an iterator that pointed to an element that was
    // removed which would lead to a crash. See Josuttis p 205.

    cout << "Erasing people of age 100" << endl;

    for (people_map::iterator j = people.begin(); j != people.end();) {
        if (j->second.GetAge() == 100)
        {

            people.erase(j++); // iterator is advanced before the erase occurs

        }
        else
            ++j; // advance the iterator
    } // end of erase loop


    // now display who is left
    cout << "Printing people left after erase ..." << endl;
    for_each(people.begin(), people.end(), fo);

    return 0;
} // end of main

如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。

扫码二维码加入Web技术交流群

发布评论

需要 登录 才能够评论, 你可以免费 注册 一个本站的账号。

评论(2

尘世孤行 2024-08-23 01:27:03

擦除使被擦除元素的迭代器无效。

 if (j->second.GetAge() == 100)
        {
            temp = j++;
            j--;
            people.erase(j); // iterator is advanced before the erase occurs
            j=temp;

        }

这不起作用,因为您将 temp 设置为等于 j 的旧值,因此您将继续使用无效的迭代器。后置自增的结果就是操作数的原始值。

我想你也可以这样做,它在功能上与工作代码相同,只是它不使用后增量的临时结果:

if (j->second.GetAge() == 100) {
    temp = j;
    ++j;
    people.erase(temp);
}

erase invalidates the iterator to the erased element.

 if (j->second.GetAge() == 100)
        {
            temp = j++;
            j--;
            people.erase(j); // iterator is advanced before the erase occurs
            j=temp;

        }

This doesn't work because you set temp equal to the old value of j, and hence you'll keep using the invalidated iterator. The result of post-increment is the original value of the operand.

I suppose you could also do it like this, which is functionally the same as the working code, except it doesn't use the temporary result of post-increment:

if (j->second.GetAge() == 100) {
    temp = j;
    ++j;
    people.erase(temp);
}
丑疤怪 2024-08-23 01:27:03

擦除会使迭代器无效,这就是使用后缀增量的原因。它不传递高级迭代器来擦除,它传递当前迭代器,但作为副作用前进。

people_map::iterator erase_it = j++;
people.erase(erase_it);

您的后缀增量有问题,这就是您的尝试失败的原因。

int i = 1;
int j = i++; // j == 1, i == 2

Erase invalidates your iterator, this is why postfix increment is used. It doesn't pass advanced iterator to erase, it passes current iterator but advances as a sideffect.

people_map::iterator erase_it = j++;
people.erase(erase_it);

You have a problem with postfix increment, this is why your attempts have failed.

int i = 1;
int j = i++; // j == 1, i == 2
~没有更多了~
我们使用 Cookies 和其他技术来定制您的体验包括您的登录状态等。通过阅读我们的 隐私政策 了解更多相关信息。 单击 接受 或继续使用网站,即表示您同意使用 Cookies 和您的相关数据。
原文