Unify the approach of integer bounds assertions

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.

Start by reading issue #109, then trace how `inRange` preconditions and Enrich-generated bounds are currently handled. The issue provides no files or tests; done would require an agreed unified approach that avoids duplicate assertions and preserves the intended semantics of implicit bounds.

Written by the indexing model from the issue text.