|
From: Julian S. <js...@ac...> - 2005-10-17 22:51:23
|
> You can't expect a memory checker to find mathematical proofs that a > value computed from a set of values that include underfined ones is > deterministic. You can always find more and more complex examples of > this sort of thing. Just so. It would be interesting to know if building a system which *always* correctly differentiates uninitialised vs initialised bits is equivalent to solving the halting problem (viz, undecideable). I suspect so. J |