C++Talk.NET C++ language newsgroups   Archives   FAQ   Search   Memberlist   Usergroups   Register   Profile   Log in to check your private messages   Log in  MPL: detecting if T provides nested type member?          C++Talk.NET Forum Index -> C++ Language (Moderated) View previous topic :: View next topic   Author Message Slawomir Lisznianski Guest Posted: Sun Oct 16, 2005 9:53 am    Post subject: MPL: detecting if T provides nested type member? Hello, Given these two types: struct has_return_type { typedef int return_type; }; struct lacks_return_type { }; Is there a way to distinguish between has_return_type and lacks_return_type? I'd like to assume some default type in case user doesn't provide the nested type explicitly. Cheers, -- Slawomir Lisznianski Programs for Research Machinery, LLC. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top David Abrahams Guest Posted: Sun Oct 16, 2005 11:18 pm    Post subject: Re: MPL: detecting if T provides nested type member? Slawomir Lisznianski <newsgroup (AT) paramay (DOT) com> writes: Quote: Hello, Given these two types: struct has_return_type { typedef int return_type; }; struct lacks_return_type { }; Is there a way to distinguish between has_return_type and lacks_return_type? #include <boost/mpl/has_xxx.hpp> BOOST_MPL_HAS_XXX_TRAIT_DEF(return_type) will define a unary has_return_type metafunction. Quote: I'd like to assume some default type in case user doesn't provide the nested type explicitly. I think you should look at the result_of<> metafunction; it may be relevant to what you're doing. Going with an existing protocol will make your code more interoperable. http://www.boost.org/libs/utility/utility.htm#result_of HTH, -- Dave Abrahams Boost Consulting www.boost-consulting.com [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Greg Herlihy Guest Posted: Mon Oct 17, 2005 11:52 am    Post subject: Re: MPL: detecting if T provides nested type member? Slawomir Lisznianski wrote: Quote: Hello, Given these two types: struct has_return_type { typedef int return_type; }; struct lacks_return_type { }; Is there a way to distinguish between has_return_type and lacks_return_type? I'd like to assume some default type in case user doesn't provide the nested type explicitly. Using an existing template library such as Boost's mpl would probably be the most portable and hassle-free solution. But just in case you find it more fun to write your own template metaprograms, here's a program to get you going: #include <iostream> struct has_return_type { typedef int return_type; }; struct lacks_return_type { }; enum ResultValue { no = sizeof(long), Yes = sizeof(char) }; struct TTest { template <class U> static char check( typename U::return_type ); template <class U> static long check( ... ); }; int main() { std::cout << "has_return_type typedefs "return_type": "; if (sizeof(TTest::check std::cout << "yes" << std::endl; else std::cout << "no" << std::endl; std::cout << "lacks_return_type typedefs "return_type": "; if (sizeof(TTest::check std::cout << "yes" << std::endl; else std::cout << "no" << std::endl; } Output: has_return_type typedefs "return_type": yes lacks_return_type typedefs "return_type": no Greg [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Slawomir Lisznianski Guest Posted: Wed Oct 19, 2005 8:35 am    Post subject: Re: MPL: detecting if T provides nested type member? Greg Herlihy wrote: Quote: Using an existing template library such as Boost's mpl would probably be the most portable and hassle-free solution. Yup, BOOST_MPL_HAS_XXX_TRAIT_DEF(return_type) works fine. Quote: find it more fun to write your own template metaprograms, here's a program to get you going: This solution won't work in all cases though. Consider this: struct user_type { }; struct has_return_type { typedef user_type return_type; }; struct lacks_return_type { }; Since `user_type' doesn't provide a constructor which takes an integral type, the compiler overload resolution always picks "..." and prints: has_return_type typedefs "return_type": no lacks_return_type typedefs "return_type": no which is not what we want. I was able to fix your solution by modifying the signature of your first `check' to: template <class T> static char check(void(*)(typename T::return_type)); The above implementation relies on properly working SFINAE (too much to ask?) Also, you assumed that long and char types were of different sizes, which the standard doesn't guarantee. Cheers, -- Slawomir Lisznianski Paramay Group, LLC "Programs for Research Machinery" [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Greg Herlihy Guest Posted: Fri Oct 21, 2005 9:13 am    Post subject: Re: MPL: detecting if T provides nested type member? Slawomir Lisznianski wrote: Quote: Greg Herlihy wrote: Using an existing template library such as Boost's mpl would probably be the most portable and hassle-free solution. Yup, BOOST_MPL_HAS_XXX_TRAIT_DEF(return_type) works fine. find it more fun to write your own template metaprograms, here's a program to get you going: This solution won't work in all cases though. Consider this: struct user_type { }; struct has_return_type { typedef user_type return_type; }; struct lacks_return_type { }; Since `user_type' doesn't provide a constructor which takes an integral type, the compiler overload resolution always picks "..." and prints: has_return_type typedefs "return_type": no lacks_return_type typedefs "return_type": no which is not what we want. I was able to fix your solution by modifying the signature of your first `check' to: template <class T static char check(void(*)(typename T::return_type)); The above implementation relies on properly working SFINAE (too much to ask?) I think that a function pointer is too elaborate. A plain pointer seems safer: template static char check( typename T::return_type *); in which case, the 0 parameter passed to check() is not an integral constant but the null pointer constant. Quote: Also, you assumed that long and char types were of different sizes, which the standard doesn't guarantee. I gambled that a char and a long would have different sizes on your compiler - simply because the odds were tremendously in my favor. While I could have declared a return type class, the more elaborate my example becomes the more it resembles Boost, and the harder it is to see what is going on. As I noted originally, Boost contains the portable version of this SFINAE technique, my example (over?)simplified the implementation - and at the cost of portabillity - to illustrate more clearly the essential technique being applied. It would be great if C++ offered more overt support for introspection. But until then, this SPINAE ("substition failure is not an error", in case anyone is wondering) seems like a reasonably safe technque to me. Greg [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Display posts from previous: All Posts1 Day7 Days2 Weeks1 Month3 Months6 Months1 Year Oldest FirstNewest First         C++Talk.NET Forum Index -> C++ Language (Moderated) All times are GMT Page 1 of 1   Jump to: Select a forum C/C++----------------C LanguageC Language (Moderated)C++ Language (Moderated)C++ language (comp.lang.c++)C++ language, library and standards Non-English C/C++ Newsgroups----------------C++ (German)C++ (French)  You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum Powered by phpBB © 2001, 2006 phpBB Group SEO toolkit © 2004-2006 webmedic.