Some concerns about the annotation TestableInVivo

[gulyx]

It may happen that it is required to introduce a testing class that extends the actual class subject to Invivo testing.
For example see commit: 09ed87c
This new class is needed for test-only purposes and mainly because:
1. The invivo test class does not (re)defines the public methods subject to invivo testing, as they are hinerited
2. Currently GROUCHO looks for annotated methods that are directly defined in the class, so in this cases it is mandatory to bind a parent class and not its child
3. The parent class may be abstract on some methods, so that it may create issues during the checkpointing activities.
Of course, this ad-hoc class should become useless once item 2 will be fixed.

[gulyx]

Add a "wrapper" for the considered CUT also in case the method is subject to invivo testing is Static. In order to build the Context of execution for the invivo session, Groucho curretly looks for the current instance associated with the method ... but in case of static methods such an instance does not exists.
Such a "wrapper" could only expose a new (not static) method that simply redirects the invocation to the actual (static) method which is the target of the In Vivo session.