Question

1.Please explain in detail Scott Meyers’s point: Use weak_ptr for shared_ptr like pointers that can dangle. Hint: a correct explanation should also discuss dangling raw pointer.

2. In your understanding, why weak_ptr was not named strong_ptr? Please explain in detail how weak_ptr’s functions expired() and lock() work.

Answer 1

1. weak_ptr is a smart pointer that acts like shared_ptr, but doesn’t contribute to ref counting. It can tell if a

object it points to has been destroyed or not (by using expire()) weak_ptr CANNOT be dereferenced. You have to

convert it to shared_ptr to be able to use it. Thus weak_ptr isn’t a standalone smart pointer, it is a augmentation

to shared_ptr.

The property that weak_ptr isn’t a standalone smart pointer can be reflected by how it’s constructed, usually

weak_ptr is constructed from a shared_ptr.

2. . A weak reference doesn't stop the garbage collector from cleaning it up, a strong reference does. They're

used for caching on some OSes. C has no equivalent of a weak reference, as it has no garbage collection.

lock() function : Creates a new std::shared_ptr that shares ownership of the managed object. If there is no managed object, i.e. *this is empty, then the returned shared_ptr also is empty.

expired() function : expired() is only useful to do things like "remove every expired weak pointer from a buffer". It is not useful for the purpose you put it.