Re: Signed int overflow behaviour in the security context

[Robert Dewar]

The one security related argument that makes sense to me
here is the business of writing tests for overflow that
make the wrap assumption. I can see that having a security
implication.