The default std::allocator class is stateless in C++. This means any instance of an std::allocator can deallocate memory allocated by another std::allocator instance. What is then the point of having instances of allocators to allocate memory? For instance, why is memory allocated like this: allocator<T> alloc, alloc2; T* buffer = alloc.allocate(42); alloc2.deallocate(buffer); When functions could… Read More If allocators are stateless in C++, why are functions not used to allocate memory instead?

[allocator.requirements.general]/37 Throws: allocate may throw an appropriate exception. Any limitations on "appropriate" implied elsewhere? Can a valid custom allocator just throw a double on any request? Context: implementation of a noexcept function that uses allocator, but has fallback strategy to do something if all allocations fail. >Solution : Any limitations on "appropriate" implied elsewhere? No.… Read More Allocator named requirements — exceptions

I know that it deprecated and removed only in std::allocator. And I can implement it on my own allocators. But why it is deprecated? >Solution : rebind was a clunky, pre-C++11 way of taking an allocator type for T and converting it into an allocator type for U. I say "pre-C++11" because C++11 gave us… Read More Why rebind<U>::other are deprecated in C++17 and removed in C++20?

I wan’t to be able to destroy any object without loosing memory: from: `——–+——–+——–` | item 1 | item 2 | item 3 | `——–+——–+——–` to: `——–+——–+——–` | item 1 | empty | item 3 | `——–+——–+——–` Used std::vector<T>::erase is what comes closest to it, but it doesn’t allow me to acces the the old… Read More Destroy object in std::vector