* long term feature: shared_ptr less locator and resource cache
* custom allocators and EnTT allocator-aware in general (long term feature, I don't actually need it at the moment) - see #22
* debugging tools (#60): the issue online already contains interesting tips on this, look at it
* work stealing job system (see #100) + mt scheduler based on const awareness for types
* allow to replace std:: with custom implementations
* add examples (and credits) from @alanjfs :)
* custom pools example (multi instance, tables, enable/disable, and so on...)

WIP:
* registry: deprecate remove_if_exists/remove_all, add remove/erase similar to sparse set that also remove all in case of no components
* scheduler, use any (or poly?) instead of unique_ptr
* any and the like: remove constructor that accepts reference wrapper, allow only in-place T& and entt::make_any<T>(...)
* resource, forward the id to the loader from the cache and if constexpr the call to load, update doc and describe customization points
* make it possible to create views of the type `view<T, T>`, add get by index and such, allow to register custom pools by name with the registry
* add user data to type_info
* make pools available (registry/view/group), review operator| for views
* any_vector for context variables
* make const registry::view thread safe, switch to a view<T...>{registry} model (long term goal)
* weak reference wrapper example with custom storage
* paginate pools
* headless (sparse set only) view
* write documentation for custom storages and views!!
* registry: use a poly object for pools, no more pool_data type.
* make runtime views use opaque storage and therefore return also elements.
* add exclude-only views to combine with packs
* entity-aware observer, add observer functions aside observer class
* deprecate non-owning groups in favor of owning views and view packs, introduce lazy owning views
* pagination doesn't work nicely across boundaries probably, give it a look. RO operations are fine, adding components maybe not.
* snapshot: support for range-based archives
* page size 0 -> page less mode
* add example: 64 bit ids with 32 bits reserved for users' purposes
