[boost] Re: shifted_ptr: why?

From: David Abrahams (dave_at_[hidden])
Date: 2003-10-01 13:59:46

"Rozental, Gennadiy" <gennadiy.rozental_at_[hidden]> writes:

>> "Rozental, Gennadiy" <gennadiy.rozental_at_[hidden]> writes:
>> 
>> > Hi,
>> >
>> > I looked over this submission. Even though I have several 
>> issues with 
>> > the code and implementation, my main concern here is: why?
>> >
>> > Why would we need another custom smart pointer component, while it 
>> > could be easily implemented using PBSP plus custom object generator?
>>   ^^^^^^^^^^^^^^^^^^
>> Are you sure?  IIUC, PBSP would be little help the 
>> interesting part of this design.  The smart pointer interface 
>> is the easy part.
>
> I mean that it easily fit into PBSP framework. Policy implementation
> wouldn't be more complex then current one. Plus we will have all the
> advantages of the framework.
>  
>> > Did I miss something?
>> >
>> > As it stands my vote it to reject this submission. Later on 
>> it could 
>> > be implemented as specific PBSP policy.
>> 
>> And that would make it acceptable?
>
> Yes. As another custom policy for generic framework. 

If every library for which a more-general framework could be
envisioned had to wait around for that more general framework to
appear, we'd never approve anything.

I'm not saying we should approve shifted_ptr neccessarily, but the
mere fact that it isn't a policy in a PBSP framework (which we don't
have) shouldn't be grounds for rejecting it.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com