[cfe-commits] [PATCH] Fix for PR7218, and analyzer support for calloc()

[Zhongxing Xu]

On Mon, May 31, 2010 at 6:17 AM, Jordy Rose <jediknil at belkadan.com> wrote:

>
> On Sun, 30 May 2010 14:56:02 -0700, Ted Kremenek <kremenek at apple.com>
> wrote:
> > On May 28, 2010, at 5:11 PM, Jordy Rose wrote:
> >
> >> I'm not entirely happy with the replacement of *p with p[0] in
> >> GRExprEngine, but that seems to be the last place where you can tell
> the
> >> difference between a binding to p and a binding to p[0].
> >
> > How are they semantically different?  The StoreManager cares about the
> > semantics of the operation, and not how it was written.  For that we can
> go
> > back to the AST.
>
> Well, PR7218 is about confusing a direct binding to p[0] as a direct
> binding for the entire p-region when looking up p[1], since RegionBindings
> stores them the same way. If element super regions can have direct
> bindings, and they look the same as element bindings, then it's important
> to distinguish the two.
>

This is a problem of the current BindingKey design, which is supposed to be
fixed in another "flat" store model. Fortunately this problem hasn't caused
much trouble until now.


>
> (Also, thanks for the explanation about LazyCompoundVal.)
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20100531/58c72a82/attachment.html>