I don't think this is something we've had to deal with before, so there's not a "usual" solution, but I'm sure we can come up with something that doesn't require forking for each release. How about either:
1. detecting whether we are on a release before or after vivid and adjusting the expectation based on that
-or-
2. Improving the test to look for at least a few known table names that exist in both releases rather than just look at the number of list items after calling split().
I think I like #2 better personally, but I don't think that's the only solution.
I don't think this is something we've had to deal with before, so there's not a "usual" solution, but I'm sure we can come up with something that doesn't require forking for each release. How about either:
1. detecting whether we are on a release before or after vivid and adjusting the expectation based on that
-or-
2. Improving the test to look for at least a few known table names that exist in both releases rather than just look at the number of list items after calling split().
I think I like #2 better personally, but I don't think that's the only solution.