Question

Suppose I have a vector<int> myvec and I want to loop through all of the elements in reverse. I can think of a few ways of doing this:

for (vector<int>::iterator it = myvec.end() - 1; it >= myvec.begin(); --it)
{
    // do stuff here
}

for (vector<int>::reverse_iterator rit = myvec.rbegin(); rit != myvec.rend(); ++rit)
{
    // do stuff here
}

for (int i = myvec.size() - 1; i >= 0; --i)
{
    // do stuff here
}

So my question is when should I use each? Is there a difference? I know that the first one is dangerous because if I pass in an empty vector, then myvec.end() - 1 is undefined, but are there any other hazards or inefficiencies with this?

Answer 1

The reverse_iterator version shows intent and works across all containers, regardless of their contents.

The first has the deficiency you describe. It also uses >=, which won’t work for non-random-access iterators.

The third has the problem that i is an int. It won’t be able to hold as much as size() could potentially return. Making it unsigned works (vector<int>::size_type), but then we have the same problem as solution one. (0U - 1 -> Funky terminating checks -> :|)