--- layout: fc_discuss_archives title: Message 12 from Frama-C-discuss on October 2012 ---
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Frama-c-discuss] Problem with Jessie and float division



Le lundi 08 octobre 2012 ? 15:38 +0200, Boris Hollas a ?crit :
> On 08.10.2012 11:36, Guillaume Melquiond wrote:
> > that make it a problem with jessie and float division? In fact, cvc3,
> > z3, and gappa, succeed in proving your program.
> 
> Indeed, CVC3 proves the code, even without the lemma, but only if I use 
> the why3-backend. However, Z3 runs in a time out.
> 
> I don't understand why because Z3 and Simplify generally perform well on 
> arithmetics.
> 
> I haven't installed gappa. I read 
> http://gappa.gforge.inria.fr/doc/ch06s01.html, but I don't know how to 
> install it in a way so that why can use it. In what order do gappa, 
> jessie, why3 have to be compiled and installed?

Whichever order you want. The only constraint is that gappa has to be
already installed before you launch the prover detection of why2 and
why3, obviously.

> Does gappa work with why3? Is gappa independent of Frama-C?

Yes to both questions.

Note that gappa is much weaker than your average smt provers and I
wouldn't suggest its use for general program verification. It just
happens to be good at arithmetic properties such as the one in your
example though.

Best regards,

Guillaume