Confusing behaviour around compliment for "sugared" types

[heshanpadmasiri]

We represent subtypes of function, table, and object types internally using other basic types (lists and mapping). When we calculate the complement we need to make sure the top type is the corresponding top type (ie. function, table, or object) not the top type of the desugared type (list or mapping). This is supposed to be achieved by *SubtypeComplement functions,

• tableSubtypeComplement
• objectSubtypeComplement
• functionSubtypeComplement -> Make complement op for function type explicit #1220

However, it seems we can replace the top type in each of these cases with arbitrary types and nothing (except type serialization resulting in different Bdds) breaks (ie. top type doesn't seem to have an effect)

[heshanpadmasiri]

looping in @lochana-chathura