
C++ 异常
发布时间:
1、异常的概念
异常是 C++ 中处理错误的机制。当函数遇到自身无法处理的错误时,会抛出异常给直接调用者。若直接调用者也无法处理,异常便会沿着调用链向上层传递,直至被能处理该错误的调用者捕获并处理。
2、相关关键字
关键字 | 作用 | 使用场景 | 特点 |
---|---|---|---|
throw | 抛出异常 | 当程序中出现问题,函数无法处理时 | 无,直接抛出指定类型的异常对象,异常对象类型可为内置 / 自定义类型 |
try | 界定可能抛出异常的代码范围 | 将可能会抛出异常的代码放在 try 块中 | 后面通常跟随一个或多个 catch 块,用于捕获 try 块中抛出的异常 |
catch | 捕获异常 | 在需要对异常进行处理的位置 | 可设置多个 catch 块,每个块可捕获特定类型的异常 |
3、throw / try / catch 基本语法
throw 异常对象;
try
{
// 可能抛出异常的代码
}
catch (异常类型 异常对象名)
{
// 处理异常的代码
}
4、异常的抛出和匹配原则
- (1)对象类型决定处理代码:异常通过抛出对象引发,异常对象的类型决定了应该激活哪个 catch 的处理代码。
- (2)就近匹配原则:在调用链中,被选中的 catch 处理代码是与异常对象类型匹配且离抛出异常位置最近的那一个。
- (3)异常对象拷贝机制:抛出异常对象后,会生成一个该异常对象的拷贝。这个拷贝的临时对象在被 catch 处理后销毁(其机制类似于函数的传值返回)。
- (4)通用捕获情况:catch 能够捕获任意类型的异常,不过使用该方式无法明确具体的错误信息。
- (5)派生类对象捕获例外:实际应用中,抛出和捕获的匹配原则存在例外情况——可以抛出派生类对象,使用基类的 catch 块进行捕获,这种方式在实际编程中非常实用。
4、函数调用链中异常栈展开的匹配原则
沿着调用链查找匹配 catch 子句的过程被称作栈展开。 - (1)首先检查 throw 操作是否处于 try 块内部。若处于 try 块内,则开始查找与之匹配的 catch 语句。若存在匹配的 catch 语句,程序将跳转到该 catch 块进行异常处理。 - (2)若在当前 try 块对应的所有 catch 语句中,没有找到与抛出异常相匹配的 catch 块,程序会退出当前函数栈。然后,在调用当前函数的函数栈中继续查找匹配的 catch 语句。 - (3)若沿着调用链不断查找,直至到达 main 函数栈,依然没有找到匹配的 catch 语句,程序将终止运行。 - (4)为了防止因存在未捕获的异常而导致程序直接终止,在实际编程中,通常需要在异常处理的最后添加 catch(...) {} 语句,用以捕获任意类型的异常。 - (5)当程序找到匹配的 catch 子句并完成异常处理后,会继续执行该 catch 子句后面的代码。
示例
#include <iostream>
double divide(double dividend, double divisor)
{
if (divisor == 0)
{
// 若除数为 0,抛出一个字符串类型的异常
throw "Error: Division by zero!";
}
// 若除数不为 0,进行除法运算并返回结果
return dividend / divisor;
}
int main()
{
double num1, num2;
std::cout << "请输入被除数: ";
std::cin >> num1;
std::cout << "请输入除数: ";
std::cin >> num2;
try
{
double result = divide(num1, num2);
std::cout << "结果: " << result << std::endl;
}
catch (const char* error)
{
// 捕获抛出的字符串类型异常,并输出错误信息
std::cerr << error << std::endl;
}
return 0;
}
请输入被除数:1
请输入除数:0
Error:Divesion by zero!
5、异常的重新抛出
单个 catch 块可能无法彻底处理异常。鉴于此,在 catch 块捕获到异常后,可先进行校正处理,再重新抛出异常传递给更上层函数处理。
校正处理可能涵盖 - 资源清理(如释放动态分配的内存、关闭文件句柄或网络连接等)、 - 数据修复(将数据恢复到某个已知的正确状态)、 - 日志记录(记录异常的详细信息,以便后续分析)等操作。
#include <iostream>
double Division(int a, int b)
{
if (b == 0)
{
throw "Division by zero condition!";
}
return (double)a / (double)b;
}
// 中间函数,捕获异常并进行校正处理后重新抛出
void Func()
{
int* array = new int[10];
try {
int len, time;
std::cout << "请输入两个整数(用空格分隔): ";
std::cin >> len >> time;
std::cout << "除法结果: " << Division(len, time) << std::endl;
}
catch (...)
{
// 资源清理
std::cout << " Func 函数中捕获到异常,释放数组内存: delete [] " << array << std::endl;
delete[] array;
throw;
}
std::cout << "正常情况释放数组内存: delete [] " << array << std::endl;
delete[] array;
}
// 更上层函数,调用 Func 函数
void Upper()
{
try
{
Func();
}
catch (...)
{
// 日志记录
std::cout << " Upper 函数中重新捕获到异常,输出提示信息后,再次重新抛出" << std::endl;
throw;
}
}
int main()
{
try
{
Upper();
}
catch (const char* errmsg)
{
std::cout << " main 函数中最终捕获到异常: " << errmsg << std::endl;
}
return 0;
}
请输入两个整数(用空格分隔):10
Fune 函数中捕获到异常,释放数组内存:delete [] 000002A0837054B0
Upper 函数中重新捕获到异常,进行一些额外处理后,再次重新抛出
main 函数中最终捕获到异常:Division by zero condition!
Division 函数中,"Division by zero condition!" 是一个字符串字面量,C++ 中字符串字面量的类型就是 const char* 。
const char* errmsg 声明了一个名为 errmsg 的参数,errmsg 会被初始化为抛出的异常对象,即指向 "Division by zero condition!" 字符串的指针。虽然 errmsg 是一个指针,但从功能上它可以看作是对抛出的异常对象的一种引用。通过该指针,我们可以访问和操作异常对象所代表的字符串内容。
C++ 标准库的异常体系
C++ 提供了一系列标准的异常,定义在
异常 | 描述 |
---|---|
std::exception | 该异常是所有标准 C++ 异常的父类。 |
std::bad_alloc | 该异常可以通过 new 抛出。 |
std::bad_cast | 该异常可以通过 dynamic_cast 抛出。 |
std::bad_exception | 这在处理 C++ 程序中无法预期的异常时非常有用。 |
std::bad_typeid | 该异常可以通过 typeid 抛出。 |
std::logic_error | 理论上可以通过读取代码来检测到的异常。 |
std::domain_error | 当使用了一个无效的数学域时,会抛出该异常。 |
std::invalid_argument | 当使用了无效的参数时,会抛出该异常。 |
std::length_error | 当创建了太长的 std::string 时,会抛出该异常。 |
std::out_of_range | 该异常可以通过方法抛出,例如 std::vector 和 std::bitset<>::operator。 |
std::runtime_error | 理论上不可以通过读取代码来检测到的异常。 |
std::overflow_error | 当发生数学上溢时,会抛出该异常。 |
std::range_error | 当尝试存储超出范围的值时,会抛出该异常。 |
std::underflow_error | 当发生数学下溢时,会抛出该异常。 |
自定义异常体系
C++标准库/自定义异常体系的使用总结
- 实际开发
- 标准库异常体系:一些小型项目或对异常处理要求不是特别定制化的场景中较为常用。如,使用标准库容器(如 std::vector )、智能指针等时,它们抛出的标准库异常(如 std::out_of_range )能直接处理常见错误,方便快捷,且无需额外定义异常类。同时,在一些跨团队协作项目中,如果大家遵循统一规范,使用标准库异常体系可以让代码具有更好的通用性和可维护性。
- 自定义异常体系:在大型项目、企业级开发中更为常用。很多公司会自己定义一套异常继承体系以规范异常管理。大型项目业务逻辑复杂,标准库异常难以精准描述业务层面错误。此外,在一些对代码架构和异常管理要求严格的项目中,自定义异常体系有助于规范异常抛出和捕获 。
- 个人练习:前期多使用标准库异常体系,随着能力提升和练习项目复杂度增加,自定义异常体系使用会增多。
示例
#include <iostream>
#include <string>
#include <thread>
#include <chrono>
#include <cstdlib>
#include <ctime>
class Exception
{
public:
Exception(const std::string& errmsg, int id)
: _errmsg(errmsg)
, _id(id)
{}
virtual std::string what() const
{
return _errmsg;
}
protected:
std::string _errmsg;
int _id;
};
class SqlException : public Exception
{
public:
SqlException(const std::string& errmsg, int id, const std::string& sql)
: Exception(errmsg, id)
, _sql(sql)
{}
virtual std::string what() const
{
std::string str = "SqlException:";
str += _errmsg;
str += "->";
str += _sql;
return str;
}
private:
const std::string _sql;
};
class CacheException : public Exception
{
public:
CacheException(const std::string& errmsg, int id)
: Exception(errmsg, id)
{}
virtual std::string what() const
{
std::string str = "CacheException:";
str += _errmsg;
return str;
}
};
class HttpServerException : public Exception
{
public:
HttpServerException(const std::string& errmsg, int id, const std::string& type)
: Exception(errmsg, id)
, _type(type)
{}
virtual std::string what() const
{
std::string str = "HttpServerException:";
str += _type;
str += ":";
str += _errmsg;
return str;
}
private:
const std::string _type;
};
void SQLMgr()
{
std::srand(std::time(0));
if (std::rand() % 7 == 0)
{
throw SqlException("权限不足", 100, "select * from name = '张三'");
}
//throw "xxxxxx";
}
void CacheMgr()
{
std::srand(std::time(0));
if (std::rand() % 5 == 0)
{
throw CacheException("权限不足", 100);
}
else if (std::rand() % 6 == 0)
{
throw CacheException("数据不存在", 101);
}
SQLMgr();
}
void HttpServer()
{
// ...
std::srand(std::time(0));
if (std::rand() % 3 == 0)
{
throw HttpServerException("请求资源不存在", 100, "get");
}
else if (std::rand() % 4 == 0)
{
throw HttpServerException("权限不足", 101, "post");
}
CacheMgr();
}
int main()
{
while (1)
{
std::this_thread::sleep_for(std::chrono::seconds(1));
try
{
HttpServer();
}
catch (const Exception& e) // 这里捕获父类对象就可以
{
// 多态
std::cout << e.what() << std::endl;
}
catch (...)
{
std::cout << "Unkown Exception" << std::endl;
}
}
return 0;
}
异常的优点
- 相比返回错误码方式,异常对象定义好后能清晰准确地展示错误信息,甚至包含堆栈调用信息,利于更好地定位程序的 bug。
- 返回错误码方式在函数调用链中,深层函数返回错误时,需层层返回错误,最外层才能拿到错误。而异常体系中,不管是哪个调用函数出错,抛出的异常会直接跳到 main 函数中 catch 捕获的地方,main 函数可直接处理错误。
// 传统错误码方式
int ConnnectSql()
{
// 用户名密码错误
if (...)
return 1;
// 权限不足
if (...)
return 2;
return 0;
}
int ServerStart()
{
if (int ret = ConnnectSql() < 0)
return ret;
int fd = socket();
if (fd < 0)
return errno;
return 0;
}
int main()
{
if (ServerStart() < 0)
// 处理错误
;
return 0;
}
- 很多第三方库都包含异常,如 boost、gtest、gmock 等常用库,使用它们需要使用异常。
- 部分函数使用异常更好处理,例如构造函数没有返回值,不方便使用错误码方式处理;像 T& operator[] 这样的函数,若 pos 越界,只能使用异常或者终止程序,无法通过返回值表示错误。
异常的缺点
- 异常会使程序的执行流乱跳,且非常混乱,运行时出错抛异常就会乱跳,这会增加跟踪调试和分析程序的难度。
- 异常会有一些性能开销,但在现代硬件速度较快的情况下,该影响基本可忽略不计。
- C++ 没有垃圾回收机制,资源需自己管理,有了异常容易导致内存泄漏、死锁等异常安全问题,需要使用 RAII 处理资源管理问题,学习成本较高。
- C++ 标准库的异常体系定义得不好,导致大家各自定义异常体系,比较混乱。
- 异常需规范使用,否则后果严重,随意抛异常会让外层捕获的用户难以处理。
总结
异常总体而言利大于弊,所以在工程中鼓励使用异常。
另外,面向对象的语言基本都用异常处理错误,这是大势所趋。