Add total to #let-bindings where LHS is a variable

[Baltoli]

This PR addresses an issue whereby the Booster reverts to the old Haskell backend when it encounters a rule written using #let bindings. The reason for this is that the de-sugared form of #let is expressed in terms of an intermediate #lambda function that is not marked as being total.

To address the issue, this PR adds the total attribute to the desugared intermediate lambda function for let-bindings and anonymous functions when the left-hand side of the original binding is a variable.

Points to consider when reviewing this PR:

• Is the identification of when the LHS a variable correct in its handling of semantic casts and sorting? My presumption is that it is given that the "first child of #SemanticCastTo KApply" logic is replicated elsewhere.
  • Add total to #let-bindings where LHS is a variable #3798 (comment)
• Is there a way to integration-test this change other than "it doesn't break the rest of the test suite"? At the moment I'm not sure there is an obvious way to do so, but I'm open to suggestions.

Fixes #3715