Weird pointer issue

[p1r0]

Hi

I'm having the following very rare problem with abstarct class
pointers.

I have, let't say, the abstract class A and the concrete class C that
implements A.
The problem is that when I was trying to make a collection of A objects
and did:

A * a = new C;
colectionOfAs[n] = a;

and realized that all a's in the array were pointing to the same place
in memory,
Then I tried:

A *a, *b;
a = new C;
b = new C;

and to my surprise a == b was true!

Any ideas?
isn't "new" supoused to give me a NEW pointer?

Thank in advance.

[Howard]

"p1r0" <lu***************@gmail.comwrote in message
news:11**********************@m79g2000cwm.googlegr oups.com...

Hi

I'm having the following very rare problem with abstarct class
pointers.

I have, let't say, the abstract class A and the concrete class C that
implements A.
The problem is that when I was trying to make a collection of A objects
and did:

A * a = new C;
colectionOfAs[n] = a;

and realized that all a's in the array were pointing to the same place
in memory,
Then I tried:

A *a, *b;
a = new C;
b = new C;

and to my surprise a == b was true!

(From this second example, I'll assume that n varied in the first example,
correct?)

There's no obvious reason that a and b should have the same value, from just
what you've shown. Perhaps it's simply how you're viewing/outputting the
values? If you could post a complete, compilable, but minimal example of
code that exhibits this behavior, we'll gladly look at it.

-Howard

[Victor Bazarov]

p1r0 wrote:

I'm having the following very rare problem with abstarct class
pointers.

I have, let't say, the abstract class A and the concrete class C that
implements A.
The problem is that when I was trying to make a collection of A
objects and did:

A * a = new C;
colectionOfAs[n] = a;

What's 'colectionOfAs'? How is it declared?

and realized that all a's in the array were pointing to the same place
in memory,

Huh?

Then I tried:

A *a, *b;
a = new C;
b = new C;

and to my surprise a == b was true!

Huh?

Any ideas?

I think this falls under FAQ 5.8.

isn't "new" supoused to give me a NEW pointer?

Well, yes.

V
--
Please remove capital 'A's when replying by e-mail
I do not respond to top-posted replies, please don't ask

[Frederick Gotham]

p1r0 posted:

A *a, *b;
a = new C;
b = new C;

and to my surprise a == b was true!

Perhaps that could be possible if you had:

A *p1, *p2;

p1 = new(nothrow) C;
p2 = new(nothrow) C;

if(p1 == p2)
{
if(!p1)
{
/* Null pointers, that explains something */
}
else
{
/* I'm a flawed compiler! */
}
}

--

Frederick Gotham

[p1r0]

Thank you all for the raplies!

Howard:

"If you could post a complete, compilable, but minimal example of
code that exhibits this behavior, we'll gladly look at it."

I'm currently not at home and don't have the code with me. The code is
kind of big for posting but I'll see if I can come up with a short
version of it.
Frederick Gotham:

I'm shure that a and b are not NULL

[benben]

>

A *a, *b;
a = new C;
b = new C;

and to my surprise a == b was true!

Any ideas?
isn't "new" supoused to give me a NEW pointer?

Operator new is indeed supposed to give a new pointer. There are
exceptions though, like when class C overloads the operator new...

Thank in advance.

Ben

[Ron Natalie]

benben wrote:

>
Operator new is indeed supposed to give a new pointer. There are
exceptions though, like when class C overloads the operator new...

The operator new function overload isn't an implementation of the
new operator (despite it's unfortunate name) but the memory allocator
used by that operator. While it's possible that the allocator could
return the same value twice, it's hard to see how that would be useful
(and would certainly take some collusion in the constructor and even
then if there are virtual functions things may get dicey).