Currently if the result from pyExec is a PyProxy, it gets destroyed.
This switches to using `to_js` to handle this (it is better to use
than an explicit `create_proxy` since it automatically decides whether
to create a proxy or not).
I also added `destroyIfProxy` which checks if something is a `PyProxy`
and then destroys it. Each use of `pyExec` needs to call `destroyIfProxy`
on the result after it is done with it.
79ad39260e