Remove corepkgs/config.nix

This isn't used anywhere except in the configure script of the Perl
bindings. I've changed the latter to use the C++ API's Settings object
at runtime.

Co-authored-by: Eelco Dolstra <edolstra@gmail.com>

Doesn't test much, but at least ensures that the command runs properly

Add some necessary casts in the initialisation of the store's config

Instead make a separate header with the template implementation of
`BaseSetting<T>::toJSONObj` that can be included where needed

It is apparently required for using `toJSONObject()`, which we do inside
the header file (because it's in a template).

This was accidentally working when building Nix itself (presumably because
`config.hh` was always included after `nlohman/json.hpp`) but caused a
(pretty dirty) build failure in the perl bindings package.

Work around https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80431 that was
already there in the code but was accidentally removed in the last
commits

So that it can be printed by `nix describe-stores`

Allow `-` and `.` in the RFC schemes as stated by
[RFC3986](https://tools.ietf.org/html/rfc3986#section-3.1).

Practically, this is needed so that `ssh-ng` is a valid URI scheme

When opening a store, only try the stores whose `uriSchemes()` include
the current one

Using virtual inheritance means that only the default constructors of
the parent classes will be called, which isn't what we want

Using the `*Config` class hierarchy

Rework the `Store` hierarchy so that there's now one hierarchy for the
store configs and one for the implementations (where each implementation
extends the corresponding config). So a class hierarchy like

```
StoreConfig-------->Store
    |                 |
    v                 v
SubStoreConfig----->SubStore
    |                 |
    v                 v
SubSubStoreConfig-->SubSubStore
```

(with virtual inheritance to prevent DDD).

The advantage of this architecture is that we can now introspect the configuration of a store without having to instantiate the store itself

The default value is initialized when creating the setting and unchanged
after that

Don't let it just contain the value, but also the other fields of the
setting (description, aliases, etc..)

Add a new `init()` method to the `Store` class that is supposed to
handle all the effectful initialisation needed to set-up the store.
The constructor should remain side-effect free and just initialize the
c++ data structure.

The goal behind that is that we can create “dummy” instances of each
store to query static properties about it (the parameters it accepts for
example)

Directly register the store classes rather than a function to build an
instance of them.
This gives the possibility to introspect static members of the class or
choose different ways of instantiating them.

We no longer need the `*Opt` to disambiguate.

Otherwise, we will associate fixed-output derivations with outputs that
they did indeed produce, but which had the wrong hash. That's no good.

gracefully handle old daemon versions