measure code coverage of runtime verification

[kiniry]

The following is written in Assurance.md, our assurance summary.

... we can measure source-level and binary-level code
runtime verification coverage of the system. In general, our RDE
methodology guarantees extremely high statement, block, function,
function call, branch, and condition/decision coverage of the
entire modeled system. It is commonly the case that we achieve
>98% coverage "out of the box", and then use either program
analysis or subject matter expert review to automatically generate
or hand-write the remaining <2% of unit, integration, or system
test cases. This methodology means that we need only manage,
evolve, and maintain two orders of magnitude fewer test cases than
a traditional agile or test-centric development process.

Measure the aforementioned kinds of coverage of our system.

• source level x statement coverage
• source level x block coverage
• source level x function coverage
• source level x function call coverage
• source level x branch coverage
• source level x condition/decision coverage
• binary level x statement coverage
• binary level x block coverage
• binary level x function coverage
• binary level x function call coverage
• binary level x branch coverage
• source level x condition/decision coverage

Note that we have not promised to do this in the SoW, so this is something that we can/should do if we have the hours on the project, or we should do afterwards with IR&D/MOB hours.