[PATCH] D77062: [analyzer] Improve zero assumption in CStringChecke::assumeZero

[Denys Petrov via Phabricator via cfe-commits]

ASDenysPetrov added a comment.

@steakhal

> If we get the value of `**b`, we get a //NonLoc// of type //unsigned char//.
> The dump of `**b` confirms this: `reg_$4<unsigned char Element{SymRegion{reg_$0<int * a>},0 S64b,unsigned char}>`, which is a `NonLoc` in deed.

Exactly. That's what I've been trying to explaine above!
This happens because of the casts, after which CSA stores operates with the symbol (b) as it points to `int*` (though, it really is `char**`).

> IMO we should fix the root cause of this in the Core.

I can't see a direct fix for now and feel quite unconfident in that part of code. That's why I suggested to accept this change as what makes CSA codebase resistant to current bugs in the Core, which we are not able to fix for now. We can change Summary and the name of this revision for acception without objections.

@steakhal
You told that you suppose a potential fix. It would be nice, if you share the patch to review.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D77062/new/

https://reviews.llvm.org/D77062