hoptimator-cli hangs/times-out for unbounded queries

[ryannedolan]

When you enter a query with ... LIMIT N, hoptimator-cli will wait for N results and then continue gracefully. But if the LIMIT is omitted, or if there are less than N rows available, the query will hang waiting for results. This would be fine, except Ctrl-C fails to cancel the command and the underlying Flink job. The process must be killed externally, or we must wait for a pre-determined timeout. For now, we wait for 20 seconds and then cancel the job, dropping any buffered results on the floor.

Ideally, the job should continue to run until:
a) N results are ready, and thus the iterator is closed, or
b) you hit ctrl-C.