2017-12-27 439 views
3

我真不明白,当异常发生在堆与分配的内存会发生什么:内存泄漏会发生什么?

#include <iostream> 
#include <vector> 
using namespace std; 

class Base { 
private: 
    int *a; 
public: 
    Base() { 
     // a = new int[100]; 

     throw runtime_error("err"); 
    } 

    ~Base() { 
     // delete[] a; 
    } 
}; 

int main() { 
    std::vector<Base> bases; 
    Base base; 
    bases.push_back(base); 

    std::cout << bases.size() << std::endl; 

    return 0; 
} 

好,现在,它只是空的程序抛出异常。 下一个程序,在堆分配内存:

#include <iostream> 
#include <vector> 
using namespace std; 

class Base { 
private: 
    int *a; 
public: 
    Base() { 
     a = new int[100]; 

     throw runtime_error("err"); 
    } 

    ~Base() { 
     // delete[] a; 
    } 
}; 

int main() { 
    std::vector<Base> bases; 
    Base base; 
    bases.push_back(base); 

    std::cout << bases.size() << std::endl; 

    return 0; 
} 

计划1:

==14151== HEAP SUMMARY: 
==14151==  in use at exit: 72,876 bytes in 3 blocks 
==14151== total heap usage: 4 allocs, 1 frees, 72,908 bytes allocated 
==14151== 
==14151== Searching for pointers to 3 not-freed blocks 
==14151== Checked 116,088 bytes 
==14151== 
==14151== 144 bytes in 1 blocks are possibly lost in loss record 2 of 3 
==14151== at 0x4C2DB8F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) 
==14151== by 0x4EC641F: __cxa_allocate_exception (in /usr/lib/x86_64-linux-gnu/libstdc++.so.6.0.21) 
==14151== by 0x400F44: Base::Base() (in /home/twite/CLionProjects/oop_learn/driver1) 
==14151== by 0x400E25: main (in /home/twite/CLionProjects/oop_learn/driver1) 
==14151== 
==14151== LEAK SUMMARY: 
==14151== definitely lost: 0 bytes in 0 blocks 
==14151== indirectly lost: 0 bytes in 0 blocks 
==14151==  possibly lost: 144 bytes in 1 blocks 
==14151== still reachable: 72,732 bytes in 2 blocks 

方案2:

==14171== HEAP SUMMARY: 
==14171==  in use at exit: 73,276 bytes in 4 blocks 
==14171== total heap usage: 5 allocs, 1 frees, 73,308 bytes allocated 
==14171== 
==14171== Searching for pointers to 4 not-freed blocks 
==14171== Checked 116,096 bytes 
==14171== 
==14171== 144 bytes in 1 blocks are possibly lost in loss record 2 of 4 
==14171== at 0x4C2DB8F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so) 
==14171== by 0x4EC641F: __cxa_allocate_exception (in /usr/lib/x86_64-linux-gnu/libstdc++.so.6.0.21) 
==14171== by 0x400F98: Base::Base() (in /home/twite/CLionProjects/oop_learn/driver2) 
==14171== by 0x400E65: main (in /home/twite/CLionProjects/oop_learn/driver2) 
==14171== 
==14171== LEAK SUMMARY: 
==14171== definitely lost: 0 bytes in 0 blocks 
==14171== indirectly lost: 0 bytes in 0 blocks 
==14171==  possibly lost: 144 bytes in 1 blocks 
==14171== still reachable: 73,132 bytes in 3 blocks 

所以,哪来的内存泄漏? Valgrind没有告诉我有关内存泄漏的情况,但是我看到了第二个程序中的那个。

+2

编译示例时是否关闭了优化,以确保事物不会意外优化? – dasblinkenlight

+0

我编译我的程序是这样的:g ++ -std = C++ 11 -o driver example.cpp然后运行Valgrind: valgrind --leak-check = full -v ./example – TwITe

+0

尝试添加“-O0”(大写“O”后跟零)标志到'g ++'行,然后再次运行'valgrind'。 – dasblinkenlight

回答

1

当程序终止时,操作系统清理它分配的任何内存 - 通常不是问题。内存泄漏在关机时很少成为问题(除非一些重要的析构函数没有机会运行)。如果内存泄漏是真正的问题,那么当它们在运行时发生并导致程序内存使用增长无限(最终导致资源耗尽)时。

+0

你说当程序终止时,操作系统清理它分配的任何内存。但是这个程序怎么样:https://pastebin.com/3gkhXVYG 程序终止,但Valgrind显示内存泄漏。 – TwITe

+1

@TwITe Valgrind显示泄漏,因为您没有释放所有内存 - 因此将其留给操作系统进行清理。这在技术上*是泄漏,但这是无关紧要的。 –