Create test to confirm that OperationRepo.processQueueForever keeps poll
when it should suspend the thread while waiting for the next operation.

My theory is this polling every 5 seconds means the app process may
never fully sleep, which contributes to battery drain.

We moved code to a new getNextOps() function simply to make it testable,
no logic changes in this commit.

Created a new private Mocks class to encapsulate creating the depended
mocks into a one liner each test can use. Some tests override the
provided mock behavior by calling every {} again.

Refactored OperationRepo.processQueueForever to prevent it from
continuously pulling when the queue is empty. My theory is this polling
every 5 seconds means the app process may never fully sleep, which
contributes to battery drain.

This commit makes the
"ensure processQueueForever suspends when queue is empty" pass.

We renamed waitForWake to waitForNewOperationAndExecutionInterval to
describe that it is doing two things.

Fixed the logic where enqueuing two or more operations would skip our
executionInterval logic. Added test coverage for this logic, as well
as a 2nd test to ensure flush can still skip waiting for it.

Lastly cleaned up the force state in the main processQueueForever loop,
we were able to fully encapsulate it into
waitForNewOperationAndExecutionInterval. The other part of the logic is
once we start processing the queue we want to do so until it's empty
again, which ops != null is the only checked needed in
processQueueForever to cover that.

We don't want to do any waiting when we call wake, so use trySend
instead. Also added result checking, to ensure we throw instead of
silently handling the error so we know if there is an issue at any time.