[cfe-dev] david's integer overflow stuff

John Regehr regehr at cs.utah.edu
Tue Sep 21 21:53:53 PDT 2010


What I mean to say is that the overflow checking logic is the actual 
invasive part, and this is already unified.  The handler that gets called 
seems like a detail.

John


On Tue, 21 Sep 2010, John Regehr wrote:

>> Is it possible to unify the two somehow?
>
> Hi Chris-
>
> If you mean "can a single trap handler provide both verbose error messages
> and also high-performance interposition like David wants" then I would
> expect that the answer is no.  It takes time and bloats code size to
> marshall up all the arguments that our handler wants.
>
> On the other hand, our interface is basically a superset of David's.  My
> belief is that suporting both trap handling interfaces will not be
> invasive at all, and that it is useful to support both.  The only
> difference is that when compiling a program in verbose trap mode, the
> trap-handling BBs will contain more stuff.
>
> John
> _______________________________________________
> cfe-dev mailing list
> cfe-dev at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev
>



More information about the cfe-dev mailing list