> Looking through here, it looks like the line:
> 
> localEvt _ anEvent transformedBy: (child transformedFrom: aMorph).
> 
> inside the loop is a loop invariant. Does this sound 
> reasonable?

No.

> You have to make the assumption that the same transform is used for a 
> morph and all of its children, which is safe as far as I know and can
> possibly still be handled if not.

Huh? To give you the simplest counter-example, draw three sketches in the
world, rotate each of them and then show me where the same transform is used
here.

Cheers,
  - Andreas