I ran through all of the optimizer/drizzle grammars + a couple of additional ones. No differences were reported, no crashes related to this patch were detected (if we saw a crash, both servers died).
I'm feeling comfortable with this despite the change in EXPLAIN output.../me has some negative thoughts on the labyrinthine optimizer code though.
If anyone wants additional testing, I can do an ExecutionTimeComparator run (checks execution time of two servers and reports if the diff is > user_set_threshold. We can also hack our grammars to see how much EXPLAIN output differs, but we're still getting the correct results for a fairly large range of queries.
I ran through all of the optimizer/drizzle grammars + a couple of additional ones. No differences were reported, no crashes related to this patch were detected (if we saw a crash, both servers died).
I'm feeling comfortable with this despite the change in EXPLAIN output.../me has some negative thoughts on the labyrinthine optimizer code though.
If anyone wants additional testing, I can do an ExecutionTimeCo mparator run (checks execution time of two servers and reports if the diff is > user_set_threshold. We can also hack our grammars to see how much EXPLAIN output differs, but we're still getting the correct results for a fairly large range of queries.