Unify the approach of integer bounds assertions

[Zoe Paraskevopoulou]

We need to have a more principled approach to generate arithmetic bounds for storage variables.

Right now:

• inRange assertions in preconditions might generate duplicate bounds assertions as Enrich already adds some go them
• inRange assertion can be implicit and generated automatically using type information, though we need to be careful to with the semantics of this. For example for a storage variable uint128 x , writing the following update
  x => (x + 1234) + 5678
  will implicitly add the precondition inRange(uint128, (x + 1234) + 5678) which implies that every subexpression never goes out of uint128 bound.

Issue #109 describes how an inferring-bounds analysis can be performed.