Re:A bug that doesn’t exist on x86: Exploiting an ARM-only r...
by usefulcat on Tuesday October 26, 2021 @03:21PM
(#29004998)
Whenever you use 'new', you have to decide what is going to 'own' the newly allocated thing. You'll also have to remember to call 'delete' somewhere.
Using unique_ptr/make_unique() or shared_ptr/make_shared() automates lifetime management (obviates the need for a manual 'delete') and makes the ownership policy explicit. They also have appropriately defined copying behavior. For example:
struct Foo {
// lots of stuff here ...
};
struct A {
Foo* f = new Foo;
~A() { delete f; }
};
struct B {
std::unique_ptr<Foo> f = std::make_unique<Foo>();
// no need to define a dtor; the default dtor is fine
};
For the destructor and the default constructor, compilers will generate basically identical code for both A and B above. If you try to copy a B, the compiler won't let you because unique_ptr isn't copyable. However it won't stop you from copying an A, even though as written (using the default copy ctor) that's almost certainly a mistake and will likely result in a double free in ~A().
Re:A bug that doesn’t exist on x86: Exploiting an ARM-only r...
by vitus on Tuesday October 26, 2021 @10:47PM
(#29009043)
Exactly as usefulcat points out: with modern C++, the bulk of object lifetimes should be handled with unique_ptr and/or directly on the stack, so your destructors are automatically called for you (reducing the risk of double frees or memory leaks).
unique_ptr forces you to think about your dependencies and when objects can / should be cleaned up.
As a naive practitioner of modern C++, I'd love it if you could elaborate on this.