Re: [Boost-users] variant over base and derived classes

Subject: Re: [Boost-users] variant over base and derived classes
From: Larry Evans (cppljevans_at_[hidden])
Date: 2010-12-09 06:12:09

On 12/08/10 18:14, Steven Watanabe wrote:
> AMDG
> 
> On 12/8/2010 4:04 PM, Hicham Mouline wrote:
>> How to determine the actual type of the object stored in a variant?
>>
>> struct S {
>>    virtual ~S()
>>    {}
>> };
>>
>> struct Deriv1 : public S {
>>    virtual ~Deriv1()
>>    {}
>> };
>>
>> struct Deriv2 : public S {
>>    virtual ~Deriv2()
>>    {}
>> };
>>
>> struct Deriv3 : public Deriv2 {
>>    virtual ~Deriv3()
>>    {}
>> };
>>
>> typedef boost::variant<S,Deriv1, Deriv2, Deriv3>  my_variant_t;
>>
>> Deriv3 d3;
>>
>> struct print_type: public boost::static_visitor<>  {
>>    template<typename T>
>>    void operator()(const T&) const
>>    {
>>      std::cout<<  typeid(T).name()<<std::endl;
>>    }
>> };
>>
>> const S&  s = d3;
>> my_variant_t v = s;
>> boost::apply_visitor( print_type(), v );
>>
>> This prints S.
>>
>> Is it a truncated copy of d3 that is stored v?
> 
> Yes.
> 
>> How do I construct v polymorphically so that the real object is stored?
> 
> You have to downcast before creating the variant.
> boost::variant only uses the static type of the object.

Why is downcasting needed instead of just
eliminating the implicit upcasting to const S&.
IOW, instead of:

  const S& s = d3;
  my_variant_t v = s;

why not:

  my_variant_t v = d3;

?

WARNING: not tested.

-Larry