is it safe to use this pointer inside destructor?

[Guest]

e.g.

SomeClass::~SomeClass()
{
SomeGlobalFunction(this);
}

does "this" still point to SomeClass instance?

 

[Michal Nazarewicz]

SomeClass::~SomeClass() {
SomeGlobalFunction(this);
}

does "this" still point to SomeClass instance?

Yes.

 

[Salt_Peter]

e.g.

SomeClass::~SomeClass()
{
SomeGlobalFunction(this);
}

does "this" still point to SomeClass instance?

Yes it does. Also, member destruction is invoked at the end of that
scope.

 

[Ron Natalie]

Yes it does. Also, member destruction is invoked at the end of that
scope.

But the more derived parts are no longer valid. Hence the
implementation will use SomeClass's type for any dynamic
resolution.

 

[Salt_Peter]

But the more derived parts are no longer valid. Hence the
implementation will use SomeClass's type for any dynamic
resolution.

Agreed. If the OP wants to shoot himself in the foot, he certainly can.
That global function might conceavably be called in all the d~tors
within an inheritance hierachy.
So when the Derived d~tor is invoked:
a) it invokes member destruction to completion
b) and then invokes the Base class d~tor to completion.
The this parameter in ~Base() can't access the Derived class nor the
Derived members polymorphicly.
That, in essence, is your point.

ie:
#include <iostream>
#include <ostream>

class Base;
void SomeGlobalFunction(Base* p);

class Member
{
public:
~Member() { std::cout << "~Member()\n"; }
};

class Base
{
public:
virtual ~Base()
{
std::cout << "~Base()\n";
SomeGlobalFunction(this);
}
virtual void foo() { std::cout << "Base::foo()\n"; }
};

class Derived : public Base
{
Member member;
public:
Derived() : member() { }
~Derived()
{
std::cout << "~Derived()\n";
SomeGlobalFunction(this);
}
void foo() { std::cout << "Derived::foo()\n"; }
};

void SomeGlobalFunction(Base* p)
{
p->foo();
}

int main()
{
Base *p = new Derived;
delete p;
return 0;
}

/*
~Derived()
Derived::foo()
~Member()
~Base()
Base::foo()
*/

 

[Salt_Peter]

But the more derived parts are no longer valid. Hence the
implementation will use SomeClass's type for any dynamic
resolution.

Agreed. If the OP wants to shoot himself in the foot, he certainly can.
That global function might conceavably be called in all the d~tors
within an inheritance hierachy.
So when the Derived d~tor is invoked:
a) it invokes member destruction to completion
b) and then invokes the Base class d~tor to completion.
The this parameter in ~Base() can't access the Derived class nor the
Derived members polymorphicly.
That, in essence, is your point.

ie:
#include <iostream>
#include <ostream>

class Base;
void SomeGlobalFunction(Base* p);

class Member
{
public:
~Member() { std::cout << "~Member()\n"; }
};

class Base
{
public:
virtual ~Base()
{
std::cout << "~Base()\n";
SomeGlobalFunction(this);
}
virtual void foo() { std::cout << "Base::foo()\n"; }
};

class Derived : public Base
{
Member member;
public:
Derived() : member() { }
~Derived()
{
std::cout << "~Derived()\n";
SomeGlobalFunction(this);
}
void foo() { std::cout << "Derived::foo()\n"; }
};

void SomeGlobalFunction(Base* p)
{
p->foo();
}

int main()
{
Base *p = new Derived;
delete p;
return 0;
}

/*
~Derived()
Derived::foo()
~Member()
~Base()
Base::foo()
*/