@John: I agree with Mike, it is not mutually exclusive, and can be applied later.
@Mike: have you thought about the implication that every PUT, POST and DELETE operation will list the directory twice with your patch? Once at DiskFile instantiation, and then again in the final_put() method? This might cause a general slow-down we should consider and measure. Let me know if I am not understanding the patch correctly.
@John: I agree with Mike, it is not mutually exclusive, and can be applied later.
@Mike: have you thought about the implication that every PUT, POST and DELETE operation will list the directory twice with your patch? Once at DiskFile instantiation, and then again in the final_put() method? This might cause a general slow-down we should consider and measure. Let me know if I am not understanding the patch correctly.