分段故障执行方法 [英] Segmentation fault executing method

问题描述

我编码的相关（我认为）行如下。
是什么意思是有一个群集列表。
其中一个， base ，将吸收另一个（ aborbed ）。
aborbed 群集应从列表中删除。

The relevant (I think) lines of my coding are below. What is meant is that there is a list of Clusters. One of them, base, will absorb another one (the aborbed). The aborbed cluster should be erased from the list.

我遇到的第一个问题是我需要在 base 和吸收的集群中执行其他操作，在之后循环关闭。
从我的搜索中，我找到了 =&（* li）的东西。我理解的是，我得到一个指向元素 li 指向的地址的指针，虽然我做不到 absorb = li ，因为一个是迭代器而另一个是（简单？）ponter。我很感激对此有一些解释。

First problem I encountered was that I needed to perform other operations in both base and absorbed clusters, after while cycle closes. From my searches, I found the = &(*li) stuff. What I understand is that I get a pointer to the address of the element li points to, although I cannot do absorbed = li, because one is an iterator and the other a (simple?) ponter. I'd appreciate some explanation on this.

现在，更大的问题是我在行 c-> getPoints中遇到了一个分段错误（）; 方法 joinCluster（）
我做错了什么？我应该怎么做？

Now, the bigger problem is that I get a sementation fault in the line c->getPoints(); of the method joinCluster() What am I doing wrong? What should I do an why?

我在Linux x86_64中使用g ++（GCC）4.5.2。

I'm using g++ (GCC) 4.5.2 in Linux x86_64.

Cluster * base;
Cluster * absorbed;

list<Cluster>::iterator li = clusters.begin(); 
while ( li != clusters.end() ) {   
    if (li->getId() == p2) {
        absorbed = &(*li);
        li = clusters.erase(li);
    } else if (li->getId() == p1) { 
        base = &(*li);
    }
++li;
}

base->joinCluster(absorbed);


void Cluster::joinCluster(Cluster * c)
{
    set<unsigned int> pts = c->getPoints(); 
}

set<unsigned int> Cluster::getPoints()
{
return points;
}

class Cluster {
    private:
    std::set<unsigned int> points;
    public:
    std::set<unsigned int> getPoints();
};

推荐答案

list<Cluster>::iterator li = clusters.begin(); 
while ( li != clusters.end() ) {   
    if (li->getId() == p2) {
        absorbed = &(*li);
        li = clusters.erase(li);
    } else if (li->getId() == p1) { 
        base = &(*li);
    }
++li;   // <----  Don't increment when you already deleted.
}

当您删除列表中的最后一个元素时， li 变为 cluster.end（）。然后你再次增加它并且繁荣，你已经出界了。 ++ li 应该进入 else 阻止。

When you delete then last element in the list, li becomes cluster.end(). Then you increment it again and boom, you're out of bounds. ++li should go in else block.

另请注意，当您从容器中删除 li 时，吸收会保留无效的地址。

Note also, when you erase li from the container, absorbed holds an invalid adress.

这篇关于分段故障执行方法的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！