Reference confusion

[Naren]

Hello Grp,
I find this a curious thing.

Below is a code

#include <iostream>
using namespace std;

class B
{
public:
B(){cout << "constructe d B\n";}
void f(){cout << "B Hello\n";}
~B(){cout << "Destroyed B\n";}
};

class A
{
B &obj;
public:
A(B &b):obj(b){c out << "constructe d A\n";}
void f(){ cout << "A Hello\n"; obj.f();}
~A(){cout << "Destroyed A\n";}
};

class C
{
A& obj;
public:

C(A &a):obj(a){c out << "constructe d C\n";};
~C(){cout << "Destroyed C\n";}
void f(){ cout << "C Hello\n";obj.f( );}
};

int main()
{
B b;
C *c = new C(A(b));
c->f();
delete c;
return 0;
}

The output of the above is
constructed B
constructed A
constructed C
Destroyed A
C Hello
A Hello
B Hello
Destroyed C
Destroyed B

The local object A gets constructed and destroyed But the object C which has
a reference to this object still exists .This should have been a violation I
believe.

Could anyone explain this

Thaanx in advance

Rgds,
Naren.

[Karl Heinz Buchegger]

Re: Reference confusion



Naren wrote:[color=blue]
>
> int main()
> {
> B b;
> C *c = new C(A(b));[/color]

It is here where the A object, a temporary, gets created and destroyed.
[color=blue]
> c->f();
> delete c;
> return 0;
> }
>
> The output of the above is
> constructed B
> constructed A
> constructed C
> Destroyed A
> C Hello
> A Hello
> B Hello
> Destroyed C
> Destroyed B
>
> The local object A gets constructed and destroyed,[/color]

yep.
[color=blue]
> But the object C which has
> a reference to this object still exists. This should have been a violation I
> believe.[/color]

It is undefined behaviour. Anything can happen. That includes: appears to work.
Your programs behaviour may change when things are added to it, the moon has
a different phase, at high water, etc...

If you have a reference to another object, it is up to you, the programmer,
to ensure that the object still is alive.

It's pretty much the same as in:

B* pB = new B;
A MyA( *pB );
delete pB;

// now MyA is left with a reference to a B object, which no longer exists.

--
Karl Heinz Buchegger
kbuchegg@gascad .at

[Josephine Schafer]

Re: Reference confusion


"Naren" <narendranath.t s@in.bosch.com> wrote in message
news:bjpcn0$a5e $1@ns2.fe.inter net.bosch.com.. .[color=blue]
> Hello Grp,
> I find this a curious thing.
>
> Below is a code
>
> #include <iostream>
> using namespace std;
>
> class B
> {
> public:
> B(){cout << "constructe d B\n";}
> void f(){cout << "B Hello\n";}
> ~B(){cout << "Destroyed B\n";}
> };
>
> class A
> {
> B &obj;
> public:
> A(B &b):obj(b){c out << "constructe d A\n";}
> void f(){ cout << "A Hello\n"; obj.f();}
> ~A(){cout << "Destroyed A\n";}
> };
>
> class C
> {
> A& obj;
> public:
>
> C(A &a):obj(a){c out << "constructe d C\n";};
> ~C(){cout << "Destroyed C\n";}
> void f(){ cout << "C Hello\n";obj.f( );}
> };
>
> int main()
> {
> B b;
> C *c = new C(A(b));
> c->f();
> delete c;
> return 0;
> }
>
> The output of the above is
> constructed B
> constructed A
> constructed C
> Destroyed A
> C Hello
> A Hello
> B Hello
> Destroyed C
> Destroyed B
>
> The local object A gets constructed and destroyed But the object C which[/color]
has[color=blue]
> a reference to this object still exists .This should have been a violation[/color]
I[color=blue]
> believe.
>
> Could anyone explain this
>[/color]

The code does not compile on VC 7.0 and Comeau 4.3.3 (strict mode).
[color=blue]
> C *c = new C(A(b));[/color]
Taking reference to a temporary object is not correct.
g++ 3.2 accepts it in normal mode :-(
IMHO, the code is calling for undefined behavior.

HTH,
J.Schafer