Add support for net:getaddrinfo/1,2

[fadushin]

This PR adds support for the net:getaddrinfo/1,2 family of functions.

Addresses Issue #902

These changes are made under both the "Apache 2.0" and the "GNU Lesser General
Public License 2.1 or later" license terms (dual license).

SPDX-License-Identifier: Apache-2.0 OR LGPL-2.1-or-later

[fadushin]

Open question -- is this slightly over-engineered? Here's why. This functionality is being implemented as an entirely new component, with (in the case of ESP32) a Config entry, a NIF registration macro, etc. Is this really necessary?

An alternative approach would be to draw a link-time dependency between this code and the OTP-socket code, which would pull this code into the runtime "for free". The linkage might be a little "artificial", but I would be okay with that. I guess it is a question of whether anyone would have use for this feature without the OTP socket code, and I think the answer to that is, "probably not".

Thoughts?

Also, it just occurred to me that this code might not be linking into the Pico port, so that should be investigated (and might be solved by the above proposal)

[bettio]

Further nitpicking (this might save us a lot of headache).
Macro should be rewritten as: ((num_elements) * ((element_size) + CONS_SIZE)).
If you do something like: LIST_SIZE(a_elements + b_elements, TUPLE_SIZE(3)) it will expand as: (a_elements + b_elements * (TUPLE_SIZE(3) + CONS_SIZE)).
I can assure that debugging this is very painful.

[fadushin]

I don't think it's nitpicking at all, and I should have picked up on this first, so my bad. Nice catch. Please re-review change.