zipWith implementation needs investigated

[benlesh]

We're currently doing something strange with zipWith (aka the "zip operator"), where we're essentially creating a new observable with the static zip, then creating a new "blank Observable", and setting the source property to the zip, but never setting the operator property.

This seems very unusual, and I feel there are probably bugs around this, and/or it could just be refactored to something simpler.

This apparently also exists in the raceWith operator.

[benlesh]

I see why it's doing this. And it is just another smell related to lift. I can't imagine an end user thinking they need to implement an operator from a static method like this. Related to #5431

Closing for now.