Re: Exception in Constructor
Am Sonntag, 16. September 2012 09:45:32 UTC+2 schrieb Henry:
[...]
class A {
public:
A() {
m_b = new B(3);
throw 0;
}
~A() {
std::cout << __PRETTY_FUNCTION__ << std::endl;
}
private:
B* m_b;
};
[...]
int main() {
try {
A a;
}
catch (...) {
std::cout << "caught an exception :)" << std::endl;
}
return 0;
}
Inside A::A(), I new-ed B and then threw an exception on purpose,
and obviously, A::~A() and B::~B() weren't called.
It doesn't seems to me that "if a constructor finishes by throwing
an exception, the memory associated with the object itself is
cleaned up ? there is no memory leak" is true.
I am pretty certain that I interpret it incorrectly. Can someone
elaborate?
You interpreted it incorrectly. The FAQ is talking about the
sizeof(A) bytes associated with the A-object -- not anything you
allocated dynamically yourself inside its constructor. The
members/subobjects of A are properly destructed and the memory that
would have stored the state of the A-object is properly released --
either because it lived in the automatic memory whose scope is left or
because the new operator -- considering new A() -- "noticed" the
exception and released the memory.
Since an object is only considered constructed when its constructor
finishes successfully, there is nothing to be destructed. So, the dtor
won't be called.
The error in your design is that you want your A to manage a
dynamically allocated B using a raw pointer. If you destruct a raw
pointer, nothing will happen. So, you basically picked the wrong type
(a raw pointer). Try this:
#include <memory>
...
class A
{
public:
A()
: m_b(new B(3))
{ throw 0; }
private:
std::unique_ptr<B> m_b; // owning pointer
};
int main()
{
try { new A; }
catch (...) {}
}
By the way: I deliberately used new in main just to stress that even
this dynamically allocated memory (for the A object) is properly
released.
Cheers!
SG
--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]