Frank Mori Hess:
On Wednesday 19 March 2008 08:39 am, Peter Dimov wrote:
After a brief consideration I think I agree with your approach, though. Nothing better comes to mind. But it would be nice if the deleter replacement happens only if shared_from_this has been called.
I think that could be done, but would probably require a mutex to protect against shared_from_this() getting called concurrently with ownership transfer.
I'm not sure. The intended use case is to call shared_from_this from the constructor, which is being called from: shared_ptr<X> px( new X ); I can think of contrived examples where X::X passes the shared_ptr to another thread which converts it to raw pointer and then calls shared_from_this again... but I think that we can postpone handling such cases unless and until they arise.
, and another that tests the new behavior.
I can add some stuff to shared_from_this_test.cpp that calls shared_from_this from constructors.
No, the new functionality should have its own separate test.
Regarding the get_deleter breakage, I think I can modify sp_counted_impl_pd/sp_counted_impl_pda so their get_deleter methods try to unwrap the deleter D if it is of type detail::deleter_wrapper (a class I'm currently using as the deleter for the shared_from_this pointers ).
This, too, can probably wait.