Investigate why having `ConcatenatingMessagestream` be more conservative with `signalProgress` calls breaks tests in `DistributedQueryBusSubscriptionQueryTest`

[hjohn]

When we change AbstractMessageStream to only call the callback when an element is available or the stream is completed when a signalProgress occurs, this breaks a specific test in DistributedQueryBusSubscriptionQueryTest: emittingUpdateInUnitOfWorkLifecycleRunsUpdatesOnAfterCommit.

It is unclear why this is, and so for now this test case for ConcatenatingMessagestream has been disabled: shouldNotDoCallbackIfFirstIsCompletedWithoutMessages. That test however looks correct, and so it would be good to figure out the exact issue that fixes both tests or perhaps fix a bad assumption in the emittingUpdateInUnitOfWorkLifecycleRunsUpdatesOnAfterCommit test.

[smcvb]

@hjohn will investigate if this is still an issue with the recent AbstractMessageStream changes and pervasive usage.