This one made me think:
class X;
void foo(X* p)
{
delete p;
}
How can we possibly delete p
if we do not even know whether X
has visible destructor? g++ 4.5.1 gives three warnings:
warning: possible problem detected in invocation of delete operator: warning: 'p' has incomplete type warning: forward declaration of 'struct X'
And then it says:
note: neither the destructor nor the class-specific operator delete will be called, even if they are declared when the class is defined.
Wow... are compilers required to diagnose this situation like g++ does? Or is it undefined behavior?
From the standard [expr.delete]:
If the object being deleted has incomplete class type at the point of deletion and the complete class has a non-trivial destructor or a deallocation function, the behavior is undefined.
So, it's UB if there's nontrivial stuff to do, and it's ok if there isn't. Warnings aren't neccessary for UB.
It is undefined behavior.
However, you can make the compiler check for incomplete types, like boost:
// verify that types are complete for increased safety
template<class T> inline void checked_delete(T * x)
{
// intentionally complex - simplification causes regressions
typedef char type_must_be_complete[ sizeof(T)? 1: -1 ];
(void) sizeof(type_must_be_complete);
delete x;
}
Applying sizeof
to an incomplete type should trigger an error, and I suppose if that passes with some compiler, then an array of negative size would trigger an error.
It is undefined behaviour, and a common gotcha when implementing the pImpl pattern. To the best of my knowledge, there is simply no such thing as a warning that the compiler is required to emit. Warnings are elective; they're there because the compiler writer thought they would be useful.
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With