[cfe-dev] Bug: Unexpected dereference claim on --analyze

[Ted Kremenek]

On Dec 5, 2011, at 5:08 PM, Anna Zaks wrote:

> Thanks for the feedback. I think the best solution would be to mention the possible redundant check in the diagnostic. Please, file a bug report!

This is a place where taint analysis would also be potentially useful.  If the pointer value is influenced by a tainted source, it probably should be treated as being a more severe candidate for a real null dereference, even in the cases where we can't fully prove that a null dereference occurred.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20111205/dc50fb01/attachment.html>