boost::shared_ptr and polymorphism

Hi,

Is boost::shared_ptr polymorphic?

i.e., can i do safe downcasting using dynamic_cast?

Thanks.

re: boost::shared_ptr and polymorphism

bb wrote:Why ask? Why not just try and see FCOL? :-)

re: boost::shared_ptr and polymorphism

bb wrote:Almost. dynamic_cast itself won't work, because the compiler doesn't now
about the internals of shared_ptr. But you can use dymamic_pointer_cast,
which applies dynamic_cast to the stored pointer:

#include <memory>

std::tr1::shared_ptr<Basesp0(new Derived);
std::tr1::shared_ptr<Derivedsp1 = dynamic_pointer_cast<Derived>(sp0);

For more details, see section 2.8.3 of my book, "The Standard C++
Library Extensions." (std::tr1::shared_ptr is pretty much the same as
boost::shared_ptr, which it was based on)

--

-- Pete

Author of "The Standard C++ Library Extensions: a Tutorial and
Reference." For more information about this book, see
www.petebecker.com/tr1book.

re: boost::shared_ptr and polymorphism

On 28 Sep 2006 09:59:23 -0700, "bb" <muralibala68@gmail.comwrote:How do you come to the idea that 'smart' pointers are pointers??

re: boost::shared_ptr and polymorphism

rpbg123@yahoo.com (Roland Pibinger) writes:
That's not helpful. You don't like smart pointers, so you're taking a
beginners question to express it by putting smart in quotes and
implying that they are not proper pointers. What's the point of this?

Besides, I don't see any connection between your mockery and the
original question, given that shared_ptr does provide a means for
doing what the OP wants, namely static_pointer_cast and
dynamic_pointer_cast.

Regards,

Jens

re: boost::shared_ptr and polymorphism

On 29 Sep 2006 00:14:42 +0100, Jens Theisen <jth02@arcor.dewrote:IT's not a question of like and dislike.
The point is that smart pointers are classes/templates which mimic
pointers by overloading operator* but which are, of course, not 'real'
pointers (as definded in the C/C++ language specifications) and also
no replacement for 'real' pointers. Ignoring the difference leads to
surprises (see eg. above). BTW, this was discussed years ago in an
article that generated some attention:
http://www-sor.inria.fr/publi/SPC++_usenixC++92.html.

Best wishes,
Roland Pibinger

re: boost::shared_ptr and polymorphism

Roland Pibinger wrote:
Hm, if your position is that smart pointers are not pointers, then I wonder
why you wrote

'smart' pointers

(insinuating they ain't smart) instead of

smart 'pointers'

(hinting at them not being pointers). Maybe, to a certain degree this is a
matter of likes and dislikes.


Best

Kai-Uwe Bux

re: boost::shared_ptr and polymorphism

Thanks Peter.

Pete Becker wrote: