Mixture of Integer and Float arithmetic without "fromIntegral"?

[Christian Lescher]

> One complication is that some Integer values exceed the greatest
> Double value.  In such cases, it's not clear which implementation of
> (/) should be used.
>
> It would certainly be nice to have an automatic conversion if, for
> example, n were an Int and x were a Float, because those types are
> both subsumed by Double.

Let's say we restrict it to the types Int and Double. Is an automatic
conversion at least in this case possible?

Christian