Tests related to generalizing monadic binds and do-notation to universe polymorphic monads in Lean 4.

• Main discussion thread on the Lean 4 Zulipchat

TODO

• A lot more examples with trickier inference situations:
  • Stress on universe inference in occurrences of Prod rather than MProd
  • Complex combinators like for in do, try catch, etc
  • Polymorphic monad operations based on eg. [MonadRef m], [MonadEnv m], [MonadLift m m']
  • "Auto-lifting"
• Generalizing the output type to cover more complex bind instances
• Alternate approach: consider trying is_monadic