Back to General and Gameplay Programming

Weird C++ type conversion problem (HARD)...

Dire.Wolf · 2000-09-09T13:15:21

I am having on hell of a problem fleshing out my Handle template class. I have created a Handle template class and I would like it to support inheritance relationships: Ex: template<class T> class Handle { public: friend class T2; // constructor explicit Handle(T* p) : ptr(p), count(new long(1)) { } // copy constructor Handle(const Handle<T>& p) throw() : ptr(p.ptr), count(p.count) { ++*count; } // destructor ~Handle() throw () { release(); } // assignment operator Handle<T>& operator=(const Handle<T>& p) throw() { if(this != &p) // protect from self assignment { release(); ptr = p.ptr; count = p.count; ++*count; } return *this; } // conversion operator, used for inheritance template<class T2> operator Handle<T2>() { return Handle<T2>(p); } // return a reference via the deferencing operator T& operator *() const throw() { return *ptr; } // return a pointer to the internal data T* operator ->() const throw() { return ptr; } // assign a new representation void Bind(T* p) { release(); ptr = p; count = new long(1); } // return the reference count long Count() const throw() { return *count; } private: // decrememnt the count var and perform check to determine // if internal data reps can be freed void release() { if(--*count == 0) { delete count; delete ptr; // make sure pointers are NULL'd count = 0; ptr = 0; } } // point to rep and reference count var T *ptr; long *count; }; namespace std { using digitalfiends::Handle; template<class T> ostream& operator <<(ostream& os, const Handle<T>& rhs) { return os << *rhs; } } class Animal { public: virtual void Speak() { cout << "Animal" << endl; }; class Dog : public Animal { public: void Speak() { cout << "Dog" << endl; }; class Cat : public Animal { public: void Speak() { cout << "Cat" << endl; }; int main() { Handle<Animal> myanimal(new Animal); Handle<Dog> mydog(new Dog); myanimal = mydog; // error: no binary '=' return 0; } You can see by the conversion operator, operator Handle , that this should work but alas, it doesn't. The problem I have with this is that in Bjarne Stroustrup's, "The C++ Programming Language - Special Edition", on page 350 he shows how to make a 'smart point' that understands inheritance/polymorphism. Unfortunately it doesn't seem to work. Anyone have any other ideas or approaches? BTW I'm using Visual C++ 6.0 SP4. Thanks in advance guys, - Dire Wolf direwolf@digitalfiends.com Edited by - Dire.Wolf on 9/3/00 3:09:20 PM Edited by - Dire.Wolf on 9/3/00 3:10:46 PM

General and Gameplay Programming Programming

Started by Dire.Wolf September 03, 2000 03:08 PM

74 comments, last by Dire.Wolf 24 years, 3 months ago

vladg

122

September 08, 2000 07:13 AM

C’mon people, this thread is degenerating into a religious war. Let me quote something for y’all:

“…we should recognize that a programming language is a tool, not a solution. A good programming language -- one that helps the programmer express ideas simply, directly, and affordably -- can be a great help, but only to people who understand their application area and the limitations of their tools.”
Bjarne Stroustrup, Why no single programming language can solve every need.

Make games, not war.

vladg

122

September 08, 2000 07:16 AM

The board messed up the link, here it is in plain text:
http://www.redherring.com/mag/issue65/news-think.html

Dire.Wolf

Author

122

September 08, 2000 09:24 AM

Wow Borlands C++ 5.5 compiler is fantastic. It compiled all my template code perfectly (where-as VC choked on a few of my headers). I wish Microsoft would get their act together and bring VC up to C++ standards (maybe in Visual Studio 7.0?).

MS, in my opinion, has the best IDE and debugger around (other than a standalone debug product). Unfortunately they are obviously not very compliant...argh.

- Dire Wolf
direwolf@digitalfiends.com

[email=direwolf@digitalfiends.com]Dire Wolf[/email]
www.digitalfiends.com

Wilka

122

September 08, 2000 12:05 PM

quote:
I wish Microsoft would get their act together and bring VC up to C++ standards (maybe in Visual Studio 7.0?).

It doesn''t look like it, check out this Q&A. Specifically:

quote:
we did not have sufficient time/resources to implement partial
specialization. We did significant work on the standard library
and on improved diagnostics for template errors, one of the main
causes of frustration and hair pulling for many C++ developers
trying to use templates. We are looking very hard at what it
takes to fully support expression templates and other advanced
template-based libraries for the next release.

Also (this is 2nd hand information, so don''t blame me if it''s lies

) "I''ve had an email exchange with a MS VC++ product manager, who confirms that they have no plans to improve ISO compliance in 7.0. So any improvements are several years away at best."

null_pointer

289

September 09, 2000 12:14 PM

Oh, BTW, to declare a template class as a friend you do something like this:


template&lttypename T2> friend class smart_ptr&ltT2>

Even if that''s not exactly right, you get the idea.

- null_pointer
Sabre Multimedia

Dire.Wolf

Author

122

September 09, 2000 01:15 PM

Unfortunately MSVC++ doesn''t like that or it seems to have absolutely no effect (no errors are reported but T2 can''t access T1''s members)! Its annoying really. Borland''s C++ understands it and handles it fine.

- Dire Wolf
direwolf@digitalfiends.com

[email=direwolf@digitalfiends.com]Dire Wolf[/email]
www.digitalfiends.com

Weird C++ type conversion problem (HARD)...

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Weird C++ type conversion problem (HARD)...

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Reticulating splines