I'm working on a C project where I have an abstract class called Aggregate, which represents a container for another abstract class called Primitive.

I want to be able to iterate through an Aggregate, without worrying with the details of how the Primitive objects are actually stored.

Since I'm not really proficient with C , I have two questions:

• First, is it even possible to do something like this?
• Second, what exactly should my Aggregate class and its derived classes do in order for this to work?

Any explanation/references are very much appreciated.

CodePudding user response：

• First, is it even possible to do something like this?
• Second, what exactly should my Aggregate class and its derived classes do in order for this to work?

Yes, you "simply" add to Aggregate two virtual functions returning a begin() and an end() iterator.

struct Aggregate
{
    struct iterator { /* ... */ };
    virtual ~Aggregate() {}
    virtual iterator begin() { return {}; }
    virtual iterator end() { return begin(); }
};

You can then use range for loops and algorithms from the Standard Library for "free":

for (auto& p : aggregate) {
    p.value = 0;
}

std::copy(aggregate.begin(), aggregate.end(), aggregate_copy.begin());

You'd need some boilerplate code to implement a working iterator though, but you'll manage with a good Google search.

Working demo