Allow unresolved values in NestedChainMap

[teutoburg]

This already works in a single RecursiveNestedMapping, but the NestedChainMap will still throw an error if a recursive bang-string-key ends up pointing to a "dead end", i.e. another, unresolved, bang key.

Maybe we'd want to include a "strict mode" for this, so we can enable that for some applications in ScopeSim, where we actually want an error in such a case. Something similar is currently done in some places, i.e. check if something is still a bang key when it should be a float or whatever. So by having a "strict mode" for resolving recursive bang keys, we wouldn't need that because the NestedChainMap would just throw a KeyError. Kinda like it currently does. Hmm, this again makes me wonder if we actually want this change, or if all errors from unresolved bang keys in ScopeSim are maybe really a symptom of something else being wrong. Yeah, I'll keep this a draft for now...

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.17%. Comparing base (8fcab48) to head (7c87948).
Report is 38 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #35   +/-   ##
=======================================
  Coverage   99.16%   99.17%           
=======================================
  Files           5        5           
  Lines         361      364    +3     
=======================================
+ Hits          358      361    +3     
  Misses          3        3           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[teutoburg]

Closing in favor of #47