"Peter Dimov" <pdimov@mmltd.net> wrote in message news:<058c01c625d8$b1f98040$6507a8c0@pdimov2>...

David Maisonave wrote:

...

Please look at the current smart pointer locking method: http://code.axter.com/smart_ptr.h By using intrusive lock logic, you

...

can lock the pointee, and thereby locking all the shared_ptr objects. Here's the smart pointer destructor: inline ~smart_ptr() throw() { m_ownership_policy.destructor_lock_policy(m_type); CHECKING_POLICY::before_release(m_type); m_ownership_policy.release(m_type, m_clone_fct, m_ownership_policy); CHECKING_POLICY::after_release(m_type); }

There's similar logic in constructor and assignment operator. This should work on all three main types of reference policies, to include reference-link.

That's what I thought. But if you have support from the pointee, why not just use intrusive_ptr? It should be even faster, on all operations, not just init.

Many programs don't need thread logic, and therefore don't require a synchronize smart pointer. For those programs that do require a synchronized smart pointer, the developer may not want to add the extra logic required for intrusive_ptr, and may only want to make the minimum modifications on their target object. In any case, this policy set allows the developer to choose what's the best combination for their requirement.