Re: [boost] Re: Re: Re: Traits-based design and ODR

From: Greg Colvin (greg_at_[hidden])
Date: 2002-05-02 14:00:03

At 12:25 PM 05/02/2002, you wrote:

>"Greg Colvin" <greg_at_[hidden]> wrote in message
>news:5.1.0.14.0.20020502104307.02ad0560_at_GMMAIL...
>> >2. shared_ptr has significant size and speed overhead for COM and
>> >CORBA-style smart pointers
>>
>> In my experience, the fastest way to handle COM locally is to do AddRef
>> once when constructing a shared_ptr, and Release once when counter
>> goes to zero, and save the virtual function calls.
>
>Why do we need smart_ptr if every time I create I need to call AddRef and
>every time I destroy I need to check counter?

You automate the AddRef in a constructor or factory
function, and automate the release with a shared_ptr
deleter function.  Then shared_ptr can do a fast local
count to manage copies.  Something like:

   void com_releaser(IUnknown* p) { p->Release(); }

   template<class T> shared_ptr<T>
   make_com_shared_ptr(IUnknown* p) {
      p->AddRef();
      return shared_ptr<T> (p,com_releaser);
   }

>> >4. shared_ptr is a point in smart_ptr's design space.
>>
>> It could be, yes.  And vice versa.
>
>No.

Actually, they are more like overlapping regions than points.
If you freeze either region, you could then expand the other
region to encompass it.  But I agree that is would be easier
for the Loki design to swallow up the Boost shared_ptr design
than vice versa.

So, as, a swallowee, I want to be sure the swallower fully
digests all the swallowed nutrients.

>> Another bastion is years of experience with shared_ptr, which
>> indicates that it is "good enough" for most purposes, that it
>> can be tweaked over time to handle other important purposes,
>> and can provide hooks for user extension in unanticipated
>> ways.  The syntax is easy for the most common case, easy
>> enough for anticipated extensions, and well within the skill
>> of those who would design new extensions.
>
>I assume that if  we provide the same syntactic convenience and hooks for
>user extension we do as good as current version.

And the same binary compatibility.  Then I will be happy and
shut up, leaving the rest of you to push the boundaries of
template metaprogramming.

>Well-slept last-night, Gennadiy.

Me too.  Greg