That "works" (minus circular references) in simple cases, but it still has bad worst case performance (if your decrements chain, causing many decrements and frees).
And in something like boost::smart_ptr, the cache performance is awful because the increment/decrement is an extra indirection in a separate allocation, which is getting hit constantly.
Actually, general malloc performance is horrible and that refcounting implimentation will be the absolute worst. In an accurate GC, the objects can be moved around in memory for more efficient allocation. Add support for generations and it becomes significantly faster than manual allocation. However, refcounting misses all of that and screws up on circular referrences to boot.
"Actually," normally starts a disagreement, but AFAICT, you restated what I said and agree. Or do I misunderstand?
Sorry, you are right. It's my bad grammer getting to me again.
The purpose of the reference count is to only deallocate memory once. You create it (copy constructor, count = 1), you pass it to a function (count = 2), come out of the function (destructor, count = 1) and finally return from the original function (count = 0, deallocate). It's like passing a hashtable around in Lisp - it doesn't pass a copy, it passes a pointer, and thus is more efficient on memory usage and speed.
What about circular references?
It's a bit hard to explain, but it's a copy constructor. Not sure of all the details right now, but I think when it is passes to another function (as a parameter), the reference count goes up, and when it returns (destructor), it goes down. I don't think it can be circular, as it uses types, not pointers. The basic idea is the data stays in the same place in memory.
It uses constructors and destructors to achieve this.