C++ question: How to change base class members in functions that don't know about the derived class

I had this idea that I could make an ObjectAnimator class, that would
animate objects that were derived from an AnimatedObject class. The
ObjectAnimator keeps a list of AnimatedObjects, and called their virtual
Animate() function on a timed basis. Ok, well that works fine, but the
problem comes when I go to take the AnimatedObject off the animate list.
The AnimatedObject has a pointer in it, pAnimListNode, to a list node in the
animate list, that gets set by the ObjectAnimator when the object is put on
the animate list. But - and this is really frustrating, I must be doing
something simple wrong - as soon as execution leaves the ObjectAnimator
function which sets that pointer, the pointer somehow immediately takes on a
different value! Or maybe it is reverting to it's previous value. It's
some kind of class inheritance problem. I've tried making AnimatedObject an
indirect base, a direct base, virtual public.. nothing I do seems to fix it.
Why doesn't this work? What am I doing wrong?

If I have, say,

class GameObject : virtual public AnimatedObject {
void *pAnimListNode;
virtual void Animate();
..
..
}

...and I call pObjectAnimator->AddToAnimateList(pGameObject), inside the
fuction,
pGameObject->pAnimListNode appears to get set properly, but as soon
execution leaves the function, the value changes!

I think the problem is that ((AnimatedObject *)pGameObject)->m_pAnimListNode
is somehow different from ((GameObject *)pGameObject)->m_pAnimListNode, even
though by itself GameObject has no such member, and even though
AnimatedObject is a virtual public base class. How frustrating.

I even made a virtual function in AnimatedObject to set the pointer,
overrode that function in GameObject, and made sure that GameObject's
version of the function was getting called by
ObjectAnimator::AddToAnimateList(). It was, it set the value. The value
stuck between the call and the function, but still again, as soon as the
thing comes out of the function that sets that value, it reverts. AArgh!
I mean, I do a step and trace, and watch the value, and as soon as it comes
out, it reverts, so it's not like my program is explicitly doing something
to the value somewhere.

Any ideas on why this could be happening / how I can fix it?

~Shavais

p.s. I'm sure I can get around the problem by passing the correct value for
the pointer out of the function and Then setting the value, from a calling
function where the derived class is known. But that defeats the whole
concept. Why am I even using OOP if I'm going to do that? I should be able
to set Lots of base class members in functions where the derived class isn't
known, shouldn't I? Without the calling environment knowing or caring about
it. What do I have to do, something like...

cBaseClassManipulator.ManipulatingFunction( pMyObject ,
pMyObject->BaseClassStructure )

... every time? Where BaseClassStructure is a member of the base class of
pMyObject that cBaseClassManipulator knows about? How silly is that!?
Sorry. Anyway, any help would sure be great.

oops actually it's not a direct base class, it's a base class of the base
class of the base class. But still. Inheritance is supposed to work,
right? lol

BTW yep, I proved my fear is true:

After calling BaseClassManipulator.BaseClassMemberChanger(pObjec t),

if I look at ((BaseClass *)pObject)->m_nValue,

I get the correct result. And the other members of ((BaseClass *)pObject)
look fine, too. But if I look at

((DerivedClass *)pObject)->m_nValue,

I get an incorrect result! Even though pObject was created as a
DerivedClass object. Not only that, but after calling
BassClassMemberChanger(), other members of ((DerivedClass *)pObject) are now
screwed up too. Oh what fun. Well, so much for my grand plans.

~Shavais


"Shavais" <shavais@hotmail.com> wrote in message
news:1Pbfb.206286$mp.128903@rwcrnsc51.ops.asp.att. net...
> I had this idea that I could make an ObjectAnimator class, that would
> animate objects that were derived from an AnimatedObject class. The
> ObjectAnimator keeps a list of AnimatedObjects, and called their virtual
> Animate() function on a timed basis. Ok, well that works fine, but the
> problem comes when I go to take the AnimatedObject off the animate list.
> The AnimatedObject has a pointer in it, pAnimListNode, to a list node in
the
> animate list, that gets set by the ObjectAnimator when the object is put
on
> the animate list. But - and this is really frustrating, I must be doing
> something simple wrong - as soon as execution leaves the ObjectAnimator
> function which sets that pointer, the pointer somehow immediately takes on
a
> different value! Or maybe it is reverting to it's previous value. It's
> some kind of class inheritance problem. I've tried making AnimatedObject
an
> indirect base, a direct base, virtual public.. nothing I do seems to fix
it.
> Why doesn't this work? What am I doing wrong?
>
> If I have, say,
>
> class GameObject : virtual public AnimatedObject {
> void *pAnimListNode;
> virtual void Animate();
> .
> .
> }
>
> ..and I call pObjectAnimator->AddToAnimateList(pGameObject), inside the
> fuction,
> pGameObject->pAnimListNode appears to get set properly, but as soon
> execution leaves the function, the value changes!
>
> I think the problem is that ((AnimatedObject
*)pGameObject)->m_pAnimListNode
> is somehow different from ((GameObject *)pGameObject)->m_pAnimListNode,
even
> though by itself GameObject has no such member, and even though
> AnimatedObject is a virtual public base class. How frustrating.
>
> I even made a virtual function in AnimatedObject to set the pointer,
> overrode that function in GameObject, and made sure that GameObject's
> version of the function was getting called by
> ObjectAnimator::AddToAnimateList(). It was, it set the value. The value
> stuck between the call and the function, but still again, as soon as the
> thing comes out of the function that sets that value, it reverts. AArgh!
> I mean, I do a step and trace, and watch the value, and as soon as it
comes
> out, it reverts, so it's not like my program is explicitly doing something
> to the value somewhere.
>
> Any ideas on why this could be happening / how I can fix it?
>
> ~Shavais
>
> p.s. I'm sure I can get around the problem by passing the correct value
for
> the pointer out of the function and Then setting the value, from a calling
> function where the derived class is known. But that defeats the whole
> concept. Why am I even using OOP if I'm going to do that? I should be
able
> to set Lots of base class members in functions where the derived class
isn't
> known, shouldn't I? Without the calling environment knowing or caring
about
> it. What do I have to do, something like...
>
> cBaseClassManipulator.ManipulatingFunction( pMyObject ,
> pMyObject->BaseClassStructure )
>
> .. every time? Where BaseClassStructure is a member of the base class of
> pMyObject that cBaseClassManipulator knows about? How silly is that!?
> Sorry. Anyway, any help would sure be great.
>
> oops actually it's not a direct base class, it's a base class of the base
> class of the base class. But still. Inheritance is supposed to work,
> right? lol
>
>
>
>

Ok I finally figured this out. This is why they make data members private,
and make all those get and set functions. Seems like a lot of trouble to go
to to me, but it does appear to fix / prevent this problem.

"Shavais" <shavais@hotmail.com> wrote in message
news:oRdfb.395819$2x.133866@rwcrnsc52.ops.asp.att. net...
> BTW yep, I proved my fear is true:
>
> After calling BaseClassManipulator.BaseClassMemberChanger(pObjec t),
>
> if I look at ((BaseClass *)pObject)->m_nValue,
>
> I get the correct result. And the other members of ((BaseClass *)pObject)
> look fine, too. But if I look at
>
> ((DerivedClass *)pObject)->m_nValue,
>
> I get an incorrect result! Even though pObject was created as a
> DerivedClass object. Not only that, but after calling
> BassClassMemberChanger(), other members of ((DerivedClass *)pObject) are
now
> screwed up too. Oh what fun. Well, so much for my grand plans.
>
> ~Shavais
>
>
> "Shavais" <shavais@hotmail.com> wrote in message
> news:1Pbfb.206286$mp.128903@rwcrnsc51.ops.asp.att. net...
> > I had this idea that I could make an ObjectAnimator class, that would
> > animate objects that were derived from an AnimatedObject class. The
> > ObjectAnimator keeps a list of AnimatedObjects, and called their virtual
> > Animate() function on a timed basis. Ok, well that works fine, but the
> > problem comes when I go to take the AnimatedObject off the animate list.
> > The AnimatedObject has a pointer in it, pAnimListNode, to a list node in
> the
> > animate list, that gets set by the ObjectAnimator when the object is put
> on
> > the animate list. But - and this is really frustrating, I must be doing
> > something simple wrong - as soon as execution leaves the ObjectAnimator
> > function which sets that pointer, the pointer somehow immediately takes
on
> a
> > different value! Or maybe it is reverting to it's previous value. It's
> > some kind of class inheritance problem. I've tried making
AnimatedObject
> an
> > indirect base, a direct base, virtual public.. nothing I do seems to fix
> it.
> > Why doesn't this work? What am I doing wrong?
> >
> > If I have, say,
> >
> > class GameObject : virtual public AnimatedObject {
> > void *pAnimListNode;
> > virtual void Animate();
> > .
> > .
> > }
> >
> > ..and I call pObjectAnimator->AddToAnimateList(pGameObject), inside the
> > fuction,
> > pGameObject->pAnimListNode appears to get set properly, but as soon
> > execution leaves the function, the value changes!
> >
> > I think the problem is that ((AnimatedObject
> *)pGameObject)->m_pAnimListNode
> > is somehow different from ((GameObject *)pGameObject)->m_pAnimListNode,
> even
> > though by itself GameObject has no such member, and even though
> > AnimatedObject is a virtual public base class. How frustrating.
> >
> > I even made a virtual function in AnimatedObject to set the pointer,
> > overrode that function in GameObject, and made sure that GameObject's
> > version of the function was getting called by
> > ObjectAnimator::AddToAnimateList(). It was, it set the value. The
value
> > stuck between the call and the function, but still again, as soon as the
> > thing comes out of the function that sets that value, it reverts.
AArgh!
> > I mean, I do a step and trace, and watch the value, and as soon as it
> comes
> > out, it reverts, so it's not like my program is explicitly doing
something
> > to the value somewhere.
> >
> > Any ideas on why this could be happening / how I can fix it?
> >
> > ~Shavais
> >
> > p.s. I'm sure I can get around the problem by passing the correct value
> for
> > the pointer out of the function and Then setting the value, from a
calling
> > function where the derived class is known. But that defeats the whole
> > concept. Why am I even using OOP if I'm going to do that? I should be
> able
> > to set Lots of base class members in functions where the derived class
> isn't
> > known, shouldn't I? Without the calling environment knowing or caring
> about
> > it. What do I have to do, something like...
> >
> > cBaseClassManipulator.ManipulatingFunction( pMyObject ,
> > pMyObject->BaseClassStructure )
> >
> > .. every time? Where BaseClassStructure is a member of the base class
of
> > pMyObject that cBaseClassManipulator knows about? How silly is that!?
> > Sorry. Anyway, any help would sure be great.
> >
> > oops actually it's not a direct base class, it's a base class of the
base
> > class of the base class. But still. Inheritance is supposed to work,
> > right? lol
> >
> >
> >
> >
>
>

"Shavais" <shavais@hotmail.com> wrote in
news:oRdfb.395819$2x.133866@rwcrnsc52.ops.asp.att. net:

> BTW yep, I proved my fear is true:
>
> After calling BaseClassManipulator.BaseClassMemberChanger(pObjec t),
>
> if I look at ((BaseClass *)pObject)->m_nValue,
>
> I get the correct result. And the other members of ((BaseClass
> *)pObject) look fine, too. But if I look at
>
> ((DerivedClass *)pObject)->m_nValue,
>
> I get an incorrect result! Even though pObject was created as a
> DerivedClass object. Not only that, but after calling
> BassClassMemberChanger(), other members of ((DerivedClass *)pObject) are
> now screwed up too. Oh what fun. Well, so much for my grand plans.

I'm not sure what you're doing here. Have a code fragment:

-----------------------------
#include <iostream>

using namespace std;

class Base
{
public:
Base() { m_int = 1; }
~Base() {}

void ChangeMe() { m_int = 2; }
void ShowMe() { cout << m_int << endl; }

int m_int;
};

class Derived : public Base
{
public:
Derived() {}
~Derived() {}

void DerivedShowMe() { cout << m_int << endl; }
};

int main()
{
Derived* pDerived = new Derived;
Base* pBase = dynamic_cast<Base*>( pDerived );

pDerived->ShowMe();
pDerived->DerivedShowMe();
pBase->ShowMe();

pDerived->ChangeMe();

pDerived->ShowMe();
pDerived->DerivedShowMe();
pBase->ShowMe();
return 0;
}

-----------------------------

I think the code fragment above does what you're trying to do, and outputs
(for me) exactly what I'd expect:

1
1
1
2
2
2

Chris
--
Chris Whitworth - Programmer at Strangelite

"back when I was young, we had to travel back in time to put the tape
in so the game would load before we died."
--otama, ua2