Ranges can be UDTs with syntactic sugar for constructing them

[swernli]

Yes, I think ranges really are just UDTs with syntactic sugar for constructing them. In old Q#, you could already manually "update" any range field:

function UpdateRangeStart(r : Range, start: Int) : Range {
    return start..RangeStep(r)..RangeEnd(r);
}

function UpdateRangeStep(r : Range, step : Int) : Range {
    return RangeStart(r)..step..RangeEnd(r);
}

function UpdateRangeEnd(r : Range, end : Int) : Range {
    return RangeStart(r)..RangeStep(r)..end;
}

So I don't see any problem with w/ being sugar for those functions. In fact I'd actually recommend we refactor the compiler to literally define Range in the core library:

newtype Range = (Start : Int, Step : Int, End : Int);

and lower range expressions into constructor calls referencing the core type. That would remove all the special casing for "primitive" fields.

Originally posted by @SamarSha in #308 (comment)