Hughes, James wrote:

...

Hello all,

I want some template code to assert if the passed in template type is NOT derived from a particular base class.

I've looked through the Template Metaprogamming book, but haven't seen anything that can do this, but I seem to remember something along those lines.

Can anyone help?

James

Hi James,

the easiest way is something like the following:

#include #include

BOOST_STATIC_ASSERT (boost::is_base_of::value);

Hope this helps,

Martin

Thanks Martin - I knew I had seen it somewhere. That covers the first point, but I've just realised I need to be a bit cleverer. The incoming type can either be the type itself or a pointer to the type, or a boost::shared_ptr to the type, and the type must be derived as the original question. Depending on which it is I need to do slightly different things, so I need to, effectively, conditionally compile a bit of code depending on the results of the tests, or assert if the type (after getting rid of the pointer/shared point bit) isn't derived from my base class. Basically I'm iterating through a vector of the incoming type and calling a function in the type stored. That function comes from the base class, but in order to be really generic I need to handle the shared_ptr and pointer cases. I think I need mpl for that, but I only get to chapter three of the book before my head starts to hurt. Any ideas/hints/places I can look? James This message (including any attachments) contains confidential and/or proprietary information intended only for the addressee. Any unauthorized disclosure, copying, distribution or reliance on the contents of this information is strictly prohibited and may constitute a violation of law. If you are not the intended recipient, please notify the sender immediately by responding to this e-mail, and delete the message from your system. If you have any questions about this e-mail please notify the sender immediately.

On 8/21/07, Martin Apel wrote:

Hi James,

...

...

the easiest way is something like the following:

#include #include

BOOST_STATIC_ASSERT (boost::is_base_of::value);

Hope this helps,

Martin

Thanks Martin - I knew I had seen it somewhere.

That covers the first point, but I've just realised I need to be a bit cleverer. The incoming type can either be the type itself or a pointer to the type, or a boost::shared_ptr to the type, and the type must be derived as the original question. Depending on which it is I need to do slightly different things, so I need to, effectively, conditionally compile a bit of code depending on the results of the tests, or assert if the type (after getting rid of the pointer/shared point bit) isn't derived from my base class.

Basically I'm iterating through a vector of the incoming type and calling a function in the type stored. That function comes from the base class, but in order to be really generic I need to handle the shared_ptr and pointer cases.

I think I need mpl for that, but I only get to chapter three of the book before my head starts to hurt. Any ideas/hints/places I can look?

I'm not sure if you really need mpl to do this. To me it looks as if you need one template and multiple partial specialization, something like the following:

template<typename T> class Handler { public: void doIt (T val) { .... // default implementation } };

template<typename T> class Handler { public: void doIt (boost::shared_ptr<T> val) { ... // Impl for shared_ptr }

template <typename T> class Handler ....

Hope this helps,

Martin

_______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

Actually you can use only one Handler class and some templated function overloads: template<class Base> class Hanlder { template<class Derived> void do_it(T const& t, typename boost::enable_if< boost::is_base_of >::type*=0) { //do smth. } template<class Derived> void do_it(boost::shared_ptr<T> t, typename boost::enable_if< boost::is_base_of >::type*=0) { //do smth. } .... };

Actually you can use only one Handler class and some templated function overloads: template<class Base> class Hanlder { template<class Derived> void do_it(T const& t, typename boost::enable_if< boost::is_base_of >::type*=0) { //do smth. } template<class Derived> void do_it(boost::shared_ptr<T> t, typename boost::enable_if< boost::is_base_of >::type*=0) { //do smth. } .... }; Thanks Martin and Ovanes, I think you have put me on the right track there. Much appreciated. James Just a quick addendum to this. I tried out a combination of the two - using a BOOST_STATIC_ASSERT((boost::is_base_of >::value)) inside the do_it function described above, and found what I thought was odd behaviour with is_base_of. The following excerpt shows the behaviour class A {}; class B : public A {}; typedef A C; BOOST_STATIC_ASSERT((boost::is_base_of >::value)) // this passes BOOST_STATIC_ASSERT((boost::is_base_of >::value)) // this fails I would expect both to pass, as is_base_of should return true when both parameters are the same, and I thought that typedef did make classes the same/equivalent http://www.boost.org/doc/html/boost_typetraits/reference.html#boost_type traits.is_base_of Is there something I am missing about typedef? Or is boost::is_base_of unable to handles the aliasing? James This message (including any attachments) contains confidential and/or proprietary information intended only for the addressee. Any unauthorized disclosure, copying, distribution or reliance on the contents of this information is strictly prohibited and may constitute a violation of law. If you are not the intended recipient, please notify the sender immediately by responding to this e-mail, and delete the message from your system. If you have any questions about this e-mail please notify the sender immediately.

-----Original Message----- From: boost-users-bounces@lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Martin Apel Sent: 21 August 2007 15:00 To: boost-users@lists.boost.org Subject: Re: [Boost-users] [mpl] Is it possible to mpl_assert dependingonthe derivation of a class

Hughes, James wrote:

...

Just a quick addendum to this. I tried out a combination of the two - using a BOOST_STATIC_ASSERT((boost::is_base_of >::value)) inside the do_it function described above, and found what I thought was odd behaviour with is_base_of.

The following excerpt shows the behaviour

class A {};

class B : public A {};

typedef A C;

BOOST_STATIC_ASSERT((boost::is_base_of >::value))

// this

...

passes

BOOST_STATIC_ASSERT((boost::is_base_of >::value))

// this

...

fails

I would expect both to pass, as is_base_of should

return true when

...

both parameters are the same, and I thought that

typedef did make

...

classes the same/equivalent

http://www.boost.org/doc/html/boost_typetraits/reference.html#boost_ty

...

petraits.is_base_of

Is there something I am missing about typedef? Or is boost::is_base_of unable to handles the aliasing?

James

Hi James,

I tried this with GCC 4.2 and 3.4.6 under Linux. It passes for both cases. So it might be your compiler, which is doing something wrong.

Martin

_______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

I'm using gcc 3.4.3 (Cross compiler to Arm). I had a thought it might be down to having some pure virtual functions in A, which meant that my typedef resulting in an incomplete type, but I removed those as a test and still got the same error. But that wouldn't explain why changing from a typedef to a derivation fixed the problem. Using boost 1.33.1 btw. I'll assume it's the compiler for now!! James This message (including any attachments) contains confidential and/or proprietary information intended only for the addressee. Any unauthorized disclosure, copying, distribution or reliance on the contents of this information is strictly prohibited and may constitute a violation of law. If you are not the intended recipient, please notify the sender immediately by responding to this e-mail, and delete the message from your system. If you have any questions about this e-mail please notify the sender immediately.