Derived and base class question

[Jeroen]

Hi,

I have 2 questions that I could not find an answer for yet. Suppose I
have the following code (only illustrative...):

class A {
A(int parm = 0);
~A();
};

class B : public A {
B();
~B();
};

Question1: If I create an object of type B, then the ctor is
automatically called for A without passing arguments. How (where) do I
change the code such that I can give a parameter to the ctor of A when
creating an object of type B?

Question 2: If I have a vector:

std::vector<A*A_pointers;

then am I allowed to store pointers to B in it also?

Thanks,

Jeroen

[Stefan Naewe]

On 5/23/2007 11:44 AM, Jeroen wrote:

Hi,

I have 2 questions that I could not find an answer for yet. Suppose I
have the following code (only illustrative...):

class A {
A(int parm = 0);
~A();

virtual ~A();

};

class B : public A {
B();

B(int parm) : A(parm) { }

~B();
};

Question1: If I create an object of type B, then the ctor is
automatically called for A without passing arguments. How (where) do I
change the code such that I can give a parameter to the ctor of A when
creating an object of type B?

Question 2: If I have a vector:

std::vector<A*A_pointers;

then am I allowed to store pointers to B in it also?

Yes. But make the d'tor of A virtual (s. above).
S.
--
Stefan Naewe stefan dot naewe at atlas-elektronik dot com
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html

[Jeroen]

Stefan Naewe schreef:

On 5/23/2007 11:44 AM, Jeroen wrote:

>Hi,

I have 2 questions that I could not find an answer for yet. Suppose I
have the following code (only illustrative...):

class A {
A(int parm = 0);
~A();

virtual ~A();

>};

class B : public A {
B();

B(int parm) : A(parm) { }

> ~B();
};

Question1: If I create an object of type B, then the ctor is
automatically called for A without passing arguments. How (where) do I
change the code such that I can give a parameter to the ctor of A when
creating an object of type B?

Question 2: If I have a vector:

std::vector<A*A_pointers;

then am I allowed to store pointers to B in it also?

Yes. But make the d'tor of A virtual (s. above).
S.

OK, thanks. For question 1 I meant that when constructing an object B,
the ctor of A always should be called with a certain parameter. I quess
I have to achieve that with:

class B : public A {
B() : A(31515) { } // or whatever parameters I want here...
~B();
};

Can you give me a clue why the dtor of A must be virual? Is that only
'better coding practice' or mandatory by C++?

Jeroen

[Stefan Naewe]

On 5/23/2007 12:20 PM, Jeroen wrote:

Stefan Naewe schreef:

>On 5/23/2007 11:44 AM, Jeroen wrote:

>>Hi,

I have 2 questions that I could not find an answer for yet. Suppose I
have the following code (only illustrative...):

class A {
A(int parm = 0);
~A();

virtual ~A();

>>};

class B : public A {
B();

B(int parm) : A(parm) { }

>> ~B();
};

Question1: If I create an object of type B, then the ctor is
automatically called for A without passing arguments. How (where) do I
change the code such that I can give a parameter to the ctor of A when
creating an object of type B?

Question 2: If I have a vector:

std::vector<A*A_pointers;

then am I allowed to store pointers to B in it also?

Yes. But make the d'tor of A virtual (s. above).
S.

OK, thanks. For question 1 I meant that when constructing an object B,
the ctor of A always should be called with a certain parameter. I quess
I have to achieve that with:

class B : public A {
B() : A(31515) { } // or whatever parameters I want here...
~B();
};

Yes, that's the way to go (if 31515 is the desired value...)

Can you give me a clue why the dtor of A must be virual? Is that only
'better coding practice' or mandatory by C++?

http://www.parashift.com/c++-faq-lit....html#faq-20.7
S.
--
Stefan Naewe stefan dot naewe at atlas-elektronik dot com
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html

[Jeroen]

Stefan Naewe schreef:

On 5/23/2007 12:20 PM, Jeroen wrote:

>Stefan Naewe schreef:

>>On 5/23/2007 11:44 AM, Jeroen wrote:
Hi,

I have 2 questions that I could not find an answer for yet. Suppose I
have the following code (only illustrative...):

class A {
A(int parm = 0);
~A();
virtual ~A();
};

class B : public A {
B();
B(int parm) : A(parm) { }
~B();
};

Question1: If I create an object of type B, then the ctor is
automatically called for A without passing arguments. How (where) do I
change the code such that I can give a parameter to the ctor of A when
creating an object of type B?

Question 2: If I have a vector:

std::vector<A*A_pointers;

then am I allowed to store pointers to B in it also?
Yes. But make the d'tor of A virtual (s. above).
S.

OK, thanks. For question 1 I meant that when constructing an object B,
the ctor of A always should be called with a certain parameter. I quess
I have to achieve that with:

class B : public A {
B() : A(31515) { } // or whatever parameters I want here...
~B();
};

Yes, that's the way to go (if 31515 is the desired value...)

>Can you give me a clue why the dtor of A must be virual? Is that only
'better coding practice' or mandatory by C++?

http://www.parashift.com/c++-faq-lit....html#faq-20.7
S.

OK, thanks guys!

[James Kanze]

On May 23, 12:34 pm, Keith Halligan <[e-mail]:

On May 23, 11:20 am, Jeroen <[e-mail]:

Can you give me a clue why the dtor of A must be virual? Is that only
'better coding practice' or mandatory by C++?

The reason that the destructor must be virtual for A is that if you
have code like A* p =new B; then we delete p, then the destructor for
B should be called to correctly. If you do not have the destructor in
A virtual this cannot occur and A's destructor will only called and
not the objects destructor that p points to.

Maybe. If you do "delete p", and the static type of *p is not
the same as the dynamic type of *p, it is undefined behavior.
For very simple cases, like the above, you will generally end up
just not calling the destructor of the derived class, but this
is not guaranteed, and in more complicated cases, stranger
things (e.g. core dumps) really do happen.

--
James Kanze (GABI Software) email:ja*********@gmail.com
Conseils en informatique orientée objet/
Beratung in objektorientierter Datenverarbeitung
9 place Sémard, 78210 St.-Cyr-l'École, France, [phone]