I've been running this one over in my head all weekend. And to tell you the
truth, it bothers me.
HtVector and HtHeap both inherit Object. Object defines the member Copy()
as a virtual function with a return type of Object*. HtVector and HtHeap
both implement Copy() as HtVector* Copy and HtHeap* Copy respectively. To
me that breaks the C++ model for redefining methods. Sparc C seems to agree
with me as well giving me a hard error in compilation.
I have a solution I'd like to propose that would resolve the issue but might
break something else and I'd like to hear some feedback on it.
What if we made Object a template class like the following:
template <class T> class Object {
public:
Object() {}
virtual ~Object() {}
virtual int compare(const T&) const = 0;
virtual T *Copy() const { return NULL; }
};
The only problem with this is that we'd lose the method
virtual int compare(Object&) const;
Now one line of code that bothers me SEVERELY is in htString.h:109
int compare(const Object& s) const { return compare((const
String&)s); }
There is NO checking to see if Object is infact a string just a blind cast
to a String reference. THAT is dangerous to say the least.
Can anyone tell me a reason why we could not just ditch the virtual compare
method I mentioned above and switch to a template?
Thanks guys & gals.
--
David J. Graff
Broadvision/Atlanta - Principal Consultant
mailto:[EMAIL PROTECTED]
Phone: (678) 427-5412
This message is intended only for the use of the Addressee and may contain
information that is PRIVILEGED and CONFIDENTIAL. If you are not the intended
recipient, dissemination of this communication is prohibited. If you have
received this communication in error, please erase all copies of the message
and its attachments and notify us immediately.
_______________________________________________
htdig-dev mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/htdig-dev
