At 02:51 2005-10-26, Martin Bonner wrote:

----Original Message---- From: Victor A. Wagner Jr. [mailto:vawjr@rudbek.com] Sent: 26 October 2005 05:36 To: boost@lists.boost.org; boost@lists.boost.org Subject: Re: [boost] Is there any interest in a library that provides containers with virtual destructors?

...

At 10:04 2005-10-25, Andy Tompkins wrote:

...

I have read a few times that people sometimes wish that std::vector (or other containers) had a virtual destructor. [snip] This allows one to write something like:

struct person { string first_name; string last_name; }; class people : public virtual_vector<person> { public: //extra methods };

Is this useful?

unless I mis-read (or mis-understood...far more likely) a lot of books on how C++ works, I don't believe there is any reason for what _you've_ shown to need a virtual destructor.

You have mis-read or mis-understood (or you need better books). Consider: virtual_vector<person>* p = new people;

I guees I wasn't clear.. I was suggesting that virtual_vector has no need for existence.

delete p; That invokes undefined behaviour unless virtual_vector<person> has a virtual destructor.

are you suggesting that if it were class people: public std::vector<person> {.....}; as above the same would be true? not "in practice", "in theory".

In *practise*, you will (probably) get away with it. The most *likely* behaviour is that virtual_vector<>::~virtual_vector() will be called, without calling people::~people(). ... but that doesn't alter the fact that it *is* undefined behaviour, which is not good.

-- Martin Bonner Martin.Bonner@Pitechnology.com Pi Technology, Milton Hall, Ely Road, Milton, Cambridge, CB4 6WZ, ENGLAND Tel: +44 (0)1223 441434 _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

Victor A. Wagner Jr. http://rudbek.com The five most dangerous words in the English language: "There oughta be a law"