aidb
aidb copied to clipboard
Limit Engine Correctness Check
Check the limit result is a subset of the full result, current check may not work for floating point numbers due to weird serialization issues.
How can I reproduce this issue? I assume the issue originates from the float precision problem but I can't test my fix locally.
Although there are no errors currently, enhancing the code could help prevent potential problems. The related test is 'test_limit_engine'.