如果我取消注释该行,有人可以解释为什么以下内容无法编译
foo::const_iterator j = f.begin();
,但是如果我使用行foo::const_iterator j = f.cbegin();
它会编译?我正试图让这条线工作,就像我的std::vector
示例一样。
#include <vector>
struct foo {
struct node { };
node *first = nullptr, *last = nullptr;
struct base_iterator {
node* ptr;
base_iterator (node* n) : ptr(n) { }
};
struct iterator : base_iterator { using base_iterator::base_iterator; };
struct const_iterator : base_iterator { using base_iterator::base_iterator; };
iterator begin() { return iterator(first); }
const_iterator begin() const { return const_iterator(first); }
const_iterator cbegin() const { return const_iterator(first); }
};
// Test
int main() {
foo f;
foo::iterator i = f.begin();
// foo::const_iterator j = f.begin(); // Won't compile because f is not const.
// foo::const_iterator j = f.cbegin(); // Will compile fine.
std::vector<int> v;
std::vector<int>::const_iterator it = v.begin(); // Compiles even though v is not const.
}
答案 0 :(得分:3)
适用于std::vector
,因为所有标准库容器的迭代器都是设计以支持iterator
- &gt; const_iterator
次转化。它旨在模仿指针转换的工作方式。
每当您的两个迭代器是用户定义的类时,您需要显式添加它。您有两种选择:
转换构造函数:
struct iterator : base_iterator { using base_iterator::base_iterator; };
struct const_iterator : base_iterator {
using base_iterator::base_iterator;
const_iterator(const iterator& other) : base_iterator(other) {}
};
转化运营商:
struct const_iterator : base_iterator { using base_iterator::base_iterator; };
struct iterator : base_iterator {
using base_iterator::base_iterator;
operator const_iterator() const { /* ... */ }
};