Simplifying resource handling of legacy APIs with smart pointers
Smart pointers (unique_ptr, shared_ptr, and weak_ptr) are extremely useful, and it is, in general, safe to say that a programmer should always use these instead of allocating and freeing memory manually.
But what if objects cannot be allocated using the new operator and/or cannot be freed again using delete? Many legacy libraries come with their own allocation/destruction functions. It seems that this would be a problem because we learned that smart pointers rely on new and delete. If the creation and/or destruction of specific types of objects relies on specific factory functions' deleter interfaces, does this prevent us from getting the humongous benefits of smart pointers?
Not at all. In this recipe, we will see that we only need to perform very minimal customizations on smart pointers in order to let them follow specific procedures for allocation and destruction of specific objects.
How to do it...
In this section, we will...
