Don't assume all failures are the same

I've started trying to implement a SOP interface to onak, primarily so I can try to take advantage of this interoperability suite. I'm not sure how viable that is given that I don't intend to support encryption or signing, but I've created a basic tool that supports the version subcommand and returns SOP_UNSUPPORTED_OPTION (return code 37) for everything else. This gives me a 50% pass rate with the test suite. I think the tests should distinguish between some of the unsupported return indicators and the actual error that is expected.