Code review comment for lp:~jaypipes/drizzle/publisher

After looking at the tests here, I have to say that I'm in favor of leaving the Python testing as-is unless there is some really urgent need to not have it.

My reasoning:
* protoc doesn't produce perl code. We use the python-generated code to interact with the manifest and get useful, test-related data.
* The current python code is awfully clean and assertion based, which can fill in gaps that test-run doesn't cover effectively. It's also well-developed / not-half-baked and we can expand the Python test without a lot of fuss.

If the use of python is a deal-breaker, we probably could do something to make it work, but I fear that the effort involved would not be worth the pay-off (at least at present).

I will be looking at - https://bugs.launchpad.net/drizzle/+bug/570948 - bad command-line options hang test cases so that we can improve test-run, but it's just my feeling that we can move things forward by keeping the current test(s)