Use get_run_status() instead of get_run() to check status

[jkotilahti]

Previously, wait_for_results() and wait_for_compilation() used get_run() to check the status of the job. The RunResult returned by get_jobs() contains metadata which contains the original job request, so if the original job request was large, it could take a long time to send and decode the RunResult JSON even before job results are ready. This meant that, for example, wait_for_results() might end up checking the status only every 30 seconds (because of the RunResult overhead) while it was supposed to check the status once every second. With unlucky timing, this could then lead to a situation where the job has already finished on the server, but wait_for_results() returns APITimeoutError long after that.

This MR fixes the problem by using get_run_status() instead of get_run() to check the status in wait_for_results() and wait_for_compilation(), so that only RunStatus is polled, not RunResult.