It seems like we are seeing a number of basic collation based errors
getting reported. I am wondering if we can somehow leverage our
existing tests to get at least a one off run through the collation
codepaths.
Is there some way we could at least get the junit based non-collation
based tests to run through the collation code paths. Best case would
be just some override on a run to use collation based database with
a territory that matches the default collation so that results would
still be right. I don't know how hard this would be to do and not
screw up the tests that explicitly want to test different collations.
/mikem