Add 'VNA03-J Sequence of calls on a synchronized abstraction may not be atomic' detector with test cases

[PatrikScully]

Individually, the methods of synchronized abstraction are atomic. However, when called in succession, there is no guarantee that the combined operation is atomic, so a method could contain a race condition that allows one thread to modify a collection and a second thread to race in and modify the collection before the first thread completes.

See https://wiki.sei.cmu.edu/confluence/display/java/VNA03-J.+Do+not+assume+that+a+group+of+calls+to+independently+atomic+methods+is+atomic

[PatrikScully]

We run the checkers on some big projects, here are the results:

• Bt (3 combined, 7 simple),
• Jenkins (0 hits),
• MatSim-Libs (24 combined, 9 simple),
• NanoHTTPD (0 hits),
• OpenGrok (0 hits),
• Ttorrent (0 hits).

[hazendaz]

Due to size of this effort and fact we have had some missteps with false positives on releases, deferring this to 5.0.0. We need to get a proper stable 4.8.x before we move forwards with new features.