C++Talk.NET C++ language newsgroups   Archives   FAQ   Search   Memberlist   Usergroups   Register   Profile   Log in to check your private messages   Log in  virtual ~type_info();          C++Talk.NET Forum Index -> C++ Language (Moderated) View previous topic :: View next topic   Author Message Bobo Guest Posted: Wed Jan 21, 2004 9:19 am    Post subject: virtual ~type_info(); Hello all. Can anybody explay WTH std::type_info has a virtual destructor? AFAIK the constructor is private so you cannot inherit from it (who would want, anyway?). TIA. Bobo. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Francis Glassborow Guest Posted: Wed Jan 21, 2004 3:58 pm    Post subject: Re: virtual ~type_info(); In message <745c4429.0401202241.6aadf69d (AT) posting (DOT) google.com>, Bobo <pinzo (AT) correo (DOT) nu> writes Quote: Hello all. Can anybody explay WTH std::type_info has a virtual destructor? AFAIK the constructor is private so you cannot inherit from it (who would want, anyway?). Actually it is the copy ctor that is private (to inhibit copying). The Standard does not actually list any other ctors but that makes sense because type_info objects are created by the implementation. The reason that a virtual dtor is provided is to allow vendors to provided extended type_info objects that can be fully handled as references (there has to be some virtual function for such things as RTTI to work) -- Francis Glassborow ACCU Author of 'You Can Do It!' see http://www.spellen.org/youcandoit [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Daniel Krügler (nee Spang Guest Posted: Wed Jan 21, 2004 4:32 pm    Post subject: Re: virtual ~type_info(); Good morning, Bobo! Bobo schrieb: Quote: Hello all. Can anybody explay WTH std::type_info has a virtual destructor? AFAIK the constructor is private so you cannot inherit from it (who would want, anyway?). According to 18.5.1 the copy-constructor and the copy-assignment operator are private, not the default constructor of the type_info class. So basically your precondition fails, and one could derive from the class. Greetings from Bremen, Daniel [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Ben Hutchings Guest Posted: Wed Jan 21, 2004 7:48 pm    Post subject: Re: virtual ~type_info(); Bobo wrote: Quote: Hello all. Can anybody explay WTH std::type_info has a virtual destructor? AFAIK the constructor is private so you cannot inherit from it (who would want, anyway?). The copy-constructor is private but it has a default constructor too. The implementation is allowed to define a derived class and to make typeid() yield instances of this derived class. I suppose that on such an implementation one could use: static_cast<const extended_type_info &>(typeid(...)) (Yuck.) [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Daniel Krügler (nee Spang Guest Posted: Thu Jan 22, 2004 9:15 am    Post subject: Re: virtual ~type_info(); Hello, Bobo, again! "Daniel Krügler (nee Spangenberg)" schrieb: Quote: According to 18.5.1 the copy-constructor and the copy-assignment operator are private, not the default constructor of the type_info class. So basically your precondition fails, and one could derive from the class. Strictly speaking, this explanation is not enough. As you know, the default c'tor is not automatically provided, if there are other c'tors declared (as there are, see the copy c'tor). After reading it a second time I am not so sure any longer, that the current words are sufficient, because it strongly depends on the interpretation on class definitions presented in Our Holy Standard. 1.4.3 says: "For classes and class templates, the library clauses specify partial definitions. Private members (clause 11) are not specified, but each implementation shall supply them to complete the definitions according to the description in the library clauses." This makes me think, that std::type_info shall indeed not be default-constructible. I also did not find any other words in the description of type_info related to constructions other than "Since the copy constructor and assignment operator for type_info are private to the class, objects of this type cannot be copied." Since DefaultConstructable is not required, it seems to be an implementation-specific detail, whether this class has a default-constructor or not. The operator typeid returns l-values of type_info, so normal users don't know, **how** these objects have been constructed. So finally I conclude from that, that only compiler writers are capable of providing derived classes for type_info (like extended_type_info). This reminds me, that even the Loki library, which also provides its own TypeInfo class, does not derive from type_info, but just contains a pointer to a type_info object. Greetings from Bremen, Daniel [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Francis Glassborow Guest Posted: Thu Jan 22, 2004 2:51 pm    Post subject: Re: virtual ~type_info(); In message <400F7C44.518C51A4 (AT) bdal (DOT) de>, Daniel Krügler <dsp (AT) bdal (DOT) de> writes Quote: Hello, Bobo, again! "Daniel Krügler (nee Spangenberg)" schrieb: According to 18.5.1 the copy-constructor and the copy-assignment operator are private, not the default constructor of the type_info class. So basically your precondition fails, and one could derive from the class. Strictly speaking, this explanation is not enough. As you know, the default c'tor is not automatically provided, if there are other c'tors declared (as there are, see the copy c'tor). Put more simply, a default ctor is not compiler generated in the presence of ANY user declared ctor. However, how does a user create an instance of a type_info? Implicitly by defining a type. There is no other way to do so, nor should there be. Nonetheless implementors must create such instances (one per appropriate udt) and may want to provide extended type information, hence the virtual dtor so that type_info is a polymorphic type. -- Francis Glassborow ACCU Author of 'You Can Do It!' see http://www.spellen.org/youcandoit [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Bobo Guest Posted: Thu Jan 22, 2004 2:59 pm    Post subject: Re: virtual ~type_info(); Daniel Krügler (nee Spangenberg) <dsp (AT) bdal (DOT) de> wrote Quote: Good morning, Bobo! Bobo schrieb: According to 18.5.1 the copy-constructor and the copy-assignment operator are private, not the default constructor of the type_info class. So basically your precondition fails, and one could derive from the class. But AFAIK the default constructor of a class is only generated if there is not any contructor declared for the class, but type_info does have a constructor declaration. Bobo. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Bobo Guest Posted: Fri Jan 23, 2004 9:47 am    Post subject: Re: virtual ~type_info(); Ben Hutchings <do-not-spam-benh (AT) bwsint (DOT) com> wrote [snip] Quote: The copy-constructor is private but it has a default constructor too. I don't think so... see my other post... Quote: The implementation is allowed to define a derived class and to make typeid() yield instances of this derived class. I suppose that on such an implementation one could use: static_cast<const extended_type_info &>(typeid(...)) (Yuck.) Maybe, but you won't require a virtual destructor to make a static_cast. But you'll need it in order to make type_info a polymorphic class so you can write: dynamic_cast<const extended_type_info &>(typeid(...)) or even typeid(typeid(...)) == typeid(extended_type_info) // Cool! But I'm not sure if any implementation do this. I guess I could test it with the following line, using as some_type a simple type, an array, a class, etc.: std::cout << typeid(typeid(some_type)).name(); Bobo [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Ben Hutchings Guest Posted: Sat Jan 24, 2004 12:52 pm    Post subject: Re: virtual ~type_info(); Bobo wrote: Quote: Ben Hutchings <do-not-spam-benh (AT) bwsint (DOT) com> wrote in message news:<slrnc0ta19.jc.do-not-spam-benh (AT) tin (DOT) bwsint.com>... [snip] The copy-constructor is private but it has a default constructor too. I don't think so... see my other post... snip No, you're right. I had a dim and incorrect memory of some exception that allows the default constructor to be implicitly declared if there is a user-declared copy constructor and no other constructors. Since I have had no reason to do define a class like that I've not noticed that this is incorrect. [ 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.