Refactor StandbyDiff for better clarity and understandability

StandbyDiff still makes use of the logOnly property for deciding when to act upon node/property changes. The official documentation of logOnly states that it helps for

/**

• read-only traversal of the diff that has 2 properties: one is to log all
• the content changes, second is to drill down to properly level, so that
• missing binaries can be sync'ed if needed
  */

but it's use is a bit misleading. The first call to StandbyDiff is always with logOnly==false, while subsequent calls are done with logOnly==true. Implementing StandbyDiff without this mechanism would result in better clarity and maintainability.

Another minor improvement is to rename #binaryCheck methods and #readBinary to #fetchBinary and #fetchAndStoreBlob which is more appropriate to their purpose.