迭代器算术类型

Question

I see a lot of people resolve to signed int , but it seems to me like that is out of ignorance, because that causes signed/unsigned mismatches for container types who use unsigned size_type s.

Since there is no way to get the type of container for the provided iterators, I do not see a way to get the size_type of the assumed container to perform accurate arithmetic.

What type should we use when performing arithmetic operations on Iterators? Or even better yet, how could one get the underlying size_type of the container since iterators have no knowledge of their parent container?

Edit2:

What I mean by the underlying size type would be, for example:

std::vector<some_type>::size_type

for some iterator whose container type is std::vector<some_type>

Edit1:

I think C++17 is providing a solution to this problem via some container access functions:

These non-member functions provide a generic interface for containers, plain arrays, and std::initializer_list.

• (function) size (C++17) returns the size of a container or array
• (function) empty (C++17) checks whether the container is empty
• (function) data (C++17) obtains the pointer to the underlying array

More specifically size gives the declared type of the size field for the parent container.

Answer 1

My recommendation is to use std::iterator_traits<It>::difference_type ( It is iterator type), which is typically (but not mandated to be) std::ptrdiff_t . std::ptrdiff_t is an alias of a signed type .

I recommend this because of the following facts:

1. std::distance returns std::iterator_traits<It>::difference_type .
2. std::next and std::prev takes std::iterator_traits<It>::difference_type as parameter.
3. Arithmetic on bidirectional iterator may go backward.