Maint: introduce mock/experimental skeletons, now for based daemon

This is meant for multiple purposes (non-exhaustive):

• simulating scenarios that are rather hard to achieve in practice (overwhelming number of nodes or artificially capped resources)

• robustness testing (will other daemons survive intermittently, yet frequently missing responses to the based queries? to what extent do they rely on some causality relations like notify-after-write? what if a chance of continuously changed CIB contents causing troubles?)

• anything not fitting the former two categories

Another option would be to start from a full-fledged daemon, cutting it
to pieces self-contained, patchable pieces, but since in case of based,
it has grown rather organically as a monolith, it would be substantially
more hassle than start grab the most vital pieces and start with a clean
room. The aim is to keep additional functionality on top of said
skeleton, and let the additional functionality stay apart in opt-in
loosly coupled modules. We shall see how this works out.