struct Base1{
int b1[100];
};
struct Base2{
int b2;
virtual void foo() {}
};
struct Derived : public Base1, public Base2{
int d;
};
int main(){
Derived D;
Base2* p_B2 = &D;
std::cout << &D << "\n";
std::cout << p_B2 << "\n";
std::cout << sizeof( D) << "\n";
std::cout << sizeof( *p_B2) << "\n";
std::cout << &(D.b2) << "\n";
std::cout << &(p_B2->b2) << "\n";
return 0;
}
If the virtual function is commented out, the output is something like:
0x7fffffffdcb0
0x7fffffffde40
408
4
0x7fffffffde40
0x7fffffffde40
If the virtual function is there, the output is:
0x7fffffffdca0
0x7fffffffdca0
416
16
0x7fffffffdca8
0x7fffffffdca8
The size difference is the v-pointer + padding, but where is the offset? Why do those point to the same address? Why is there only 8 bytes (the v-pointer I assume) difference between D and D.b2 when there should be at least 100 int-s between them?
(Also my debugger (gdb) shows p_B2 as "Derived" type.)
The compiler is free to lay out your classes however it likes, presumably your compiler always puts virtual classes at the beginning and any non-virtual classes after that.
You can't rely on any particular ordering of classes.