Job results should be cached once job.result() is called

[nonhermitian]

What is the expected feature or enhancement?

Getting job results has to query the API each time job.result() is called. This makes code very slow, and should not have to be done more than once. Results used to be cached, and should be again.

Acceptance criteria
Job results are cached

[kt474]

Caching job results was removed ~6 months ago - the logic behind the decision was that job results can be very large and the user does not have an easy way to free up the memory (discussed in ch-performance)

[jyu00]

Like @kt474 said, we removed the caching because people were complaining about memory consumption after running many jobs. And one can easily cache results locally by assigning them to a local variable.