Re: [boost] Re: Re: Re: Re: TR2 and C++0x heads up

From: Howard Hinnant (hinnant_at_[hidden])
Date: 2005-04-26 21:49:02

On Apr 26, 2005, at 5:44 PM, Peter Dimov wrote:

> Howard Hinnant wrote:
>> On Apr 26, 2005, at 11:37 AM, Peter Dimov wrote:
>>
>>> ConstPropagation policy (with the default being No for shallow copy
>>> and Yes for deep copy).
>>
>> I've been toying with a smart pointer adaptor for this one.  I really
>> haven't fully fleshed out the thinking on it though...
>>
>> template <class SP>
>> class deep_const
>
> Yes, conversion and checking can be implemented as adaptors, too.
>
> But a really smart pointer would have the right default for const 
> propagation depending on the ownership policy. :-)

Would that be the prescient_ptr? :-)

I've got 3 ownership policies on my mind:

1.  shared
2.  unique
3.  cloned

And if anything I'm leaning towards shallow const on all of them.  
However I also believe deep const versions of all of these would 
probably be valuable as well.  And I'm especially looking at signatures 
like:

template<class T, class U>
bool operator==(const shared_ptr<T>& a, const shared_ptr<U>& b);

shared_ptr today is a shallow const animal, and I think that is good.  
But the above signature assumes deep const.  It would be nice to have a 
zero-overhead way to express and enforce that assumption.

So would it be better to code a const-policy into them all (and 
whatever other smart pointers crop up)?  Or would it be better to make 
a single adaptor (like stack is for containers), that would take any 
shallow const smart pointer and adapt it to a deep one?  Or maybe there 
is a third solution that has escaped me.  Or perhaps this is an 
academic problem and really requires no solution.

I'm not claiming to know the answer to my question.  I'm simply 
grateful I have a place to ask it.

-Howard