Related to #439
The behavior at the bounds is different because of how the inputs (reserves) are overwritten to be just at the bounds. The tests should make sure the behavior stays consistent as we get closer and closer to the bounds and it does not jump. In issue #439, there was scenarios where the invariant "jumps" because of the invariant's bound behaviors, which leads to unexpected economic pricing in the pools.
Related to #439
The behavior at the bounds is different because of how the inputs (reserves) are overwritten to be just at the bounds. The tests should make sure the behavior stays consistent as we get closer and closer to the bounds and it does not jump. In issue #439, there was scenarios where the invariant "jumps" because of the invariant's bound behaviors, which leads to unexpected economic pricing in the pools.