Forum OpenACS Development: Response to permission_p revisited

Just wanted to clarify that we didn't notice a performance difference between Oracle and PG, we never tested PG. I wanted to commit the reordering changes but decided not to when I saw that the Oracle and PG versions of permission_p were different. The Oracle version has one more check (5 instead of 4) and they are already in different order. I didn't want to mess around with this back then so we just left it alone.

If someone wants to look into which one is correct or what is going on that would be great.