Specify behavior of 'resetCheckpoint()' on Replicator restart

The intended behavior of the Replicator.resetCheckpoint call is no longer clear. Since setting replicator options is linked to the creation of the of the native replicator it is unclear what should happen in the following cases:

• Replicator.resetCheckpoint is called after getDocumentIds (which creates the c4Replicator)
• Replicator.start() is called a second time on a replicator on which resetCheckpoint was called, successfully.

The resolution to this issue is likely to affect all platforms.