If remote job raises exception, make local process raise the same exception, rather than generic RuntimeError.
Distinguish jobs that died (usually timed out, or maybe exceeded memory limit) from those that failed with a remote exception, so they can (in principle) be handled differently.
If remote job raises exception, make local process raise the same exception, rather than generic
RuntimeError.Distinguish jobs that died (usually timed out, or maybe exceeded memory limit) from those that failed with a remote exception, so they can (in principle) be handled differently.