Fix flaky testFailStuckSplitTasks unit test

[leetcode-1533]

Description

Fix flaky testFailStuckSplitTasks unit test.

1. Added wait for task executors to pick up tasks
2. Close task executor after the unit test finished.

There is a race condition between unit test thread and the task executor thread. Added checking via settableFuture in the unit test thread to ensure task executor to process splits first.

Is this change a fix, improvement, new feature, refactoring, or other?

Fix

Is this a change to the core query engine, a connector, client library, or the SPI interfaces? (be specific)

Query Engine

How would you describe this change to a non-technical end user or system administrator?

Fix flaky testFailStuckSplitTasks unit test.

Related issues, pull requests, and links

#12392
#13437

Documentation

(x) No documentation is needed.
( ) Sufficient documentation is included in this PR.
( ) Documentation PR is available with #prnumber.
( ) Documentation issue #issuenumber is filed, and can be handled later.

Release notes

(x) No release notes entries required.
( ) Release notes entries required with the following suggested text:

# Section
* Fix some things. ({issue}`issuenumber`)

[phd3]

lgtm

[arhimondr]

Also I assume it might be reasonable to have another future that would identify a split finished execution

[leetcode-1533]

The split is never finished // interrupted. It is only terminated at taskExecutor.stop().

The unit test only verified that the sqlTaskManager could identify the task with a long-running split and mark the task to the failed state.

It is because the sqlTaskManager issues failTask() for the task, which eventually issues a list of callback functions while doing the state transition of the task. In a real Trino cluster, one of which is responsible for interrupting the DriverSplitRunner(an implementation of SplitRunner), however, in the unit test, there is no such callback function. And it is another unit test that checks when doing state changes, the registered callback functions can be called successfully.

Let me know if you have a suggestion for improvement!

[arhimondr]

LGTM % comments