improve fake creation options interaction

[blairconrad]

spun off from/a precursor to #461, specifically #461 (comment), where I advocate this set of interactions:

• CallsBaseMethods - should trump prior invocations of ConfigureFake, Strict, and Wrapping
• ConfigureFake - should "stack" with prior invocations, just as
  multiple A.CallsTo do, which would include overriding earlier
  CallsBaseMethods, Strict, and Wrapping
• Implements - should "stack" with prior invocations, ever accumulating more interfaces
• Strict - should trump prior invocations of CallsBaseMethods,
  ConfigureFake, and Wrapping
• WithAdditionalAttributes - should "stack" with prior invocations, ever accumulating more attributes
• WithArgumentsForConstructor - should trump prior invocations
• Wrapping - should trump prior invocations of CallsBaseMethods,
  ConfigureFake, Strict, and Wrapping

While writing characterization tests in #465, I noticed deviations, even when just setting competing options in the A.Fake() calls:

• WithAdditionalAttributes called twice: last one in wins, rather than stacking
• ConfigureFake followed by Wrapping - uses configured, not wrapped, behaviour
• CallsBaseMethods followed by Wrapping - calls base method
• Strict followed by Wrapping - acts strict

I'd like to smooth out these deviations before extending IFakeConfigurator.

I am very close to a PR (in the "cleanup" phase) that would build upon the work in #465.

[blairconrad]

This issue has been fixed in release 2.0.0.