Frama-C-discuss mailing list archives

This page gathers the archives of the old Frama-C-discuss archives, that was hosted by Inria's gforge before its demise at the end of 2020. To search for mails newer than September 2020, please visit the page of the new mailing list on Renater.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Frama-c-discuss] hint assertions and understanding cooperation between wp and value plugin

  • Subject: [Frama-c-discuss] hint assertions and understanding cooperation between wp and value plugin
  • From: dcok at grammatech.com (David R. Cok)
  • Date: Wed, 4 May 2016 21:31:43 -0400
  • In-reply-to: <CABbVA-DZOHUA-e-OXdo45V7Hg2EB=WdoLU4Q_n6mxT+_wZKTpA@mail.gmail.com>
  • References: <87fuu82u9r.fsf@inria.fr> <A0F2DD80-E14A-4356-A8C3-0512BD7BC656@cea.fr> <5729E5E9.2090403@inria.fr> <CABbVA-DZOHUA-e-OXdo45V7Hg2EB=WdoLU4Q_n6mxT+_wZKTpA@mail.gmail.com>
Sorry: when you say "invalid in ACSL" do you actually mean it is invalid 
(because it is floating point, not real?) or do you mean it is not 
provable (and by what) or do you mean it is not a well-formed ACSL 
statement? ACSL is just a language, not a proof tool. If the first, is 
it actually the case that there are double precision numbers for which 
the assertion does not hold?

- David


On 5/4/2016 2:14 PM, Boris Yakobowski wrote:
> My mistake. I coalesced two different things when replying to Maurice :
>
> - the validity of the lemma/*@ lemma sq_double: \forall real x; x*x - 
> .2 * x + 0.01 >= 0.; */. It is indeed valid in ACSL, should be proven 
> by WP whatever the model (since, as Loïc mentioned, it involves only 
> real) , and (as all lemmas) is completely ignored by Value;
>
> - the validity status of the assert in double x1 = x*x - .2 * x + 
> 0.01; /*@ assert x1 >= 0; */, which is invalid in ACSL, but can be 
> proven by the Float model of WP.
>
> On Wed, May 4, 2016 at 2:07 PM, Claude Marché <Claude.Marche at inria.fr 
> <mailto:Claude.Marche at inria.fr>> wrote:
>
>
>     Sorry to interfere, but I don't understand the meaning of "wrong
>     in WP"
>
>     According to ACSL manual, this lemma is a statement expressed
>     purely in
>     mathematical real arithmetic, and as such it is valid. It is indeed
>     proved automatically by Z3 4.4.1.
>
>     If you want to state a similar property talking about floating-point
>     arithmetic, it should be stated differently, typically using a program
>
>     void f(double x) {
>       double y = x*x - .2 * x + 0.01;
>       //@ assert y >= 0.0;
>     }
>
>     But I guess it probably wrong because of rounding, even with a
>     precondition like \abs(x) <= 1.0
>
>     My two cents,
>
>     - Claude
>
>     Le 04/05/2016 13:54, Loïc Correnson a écrit :
>     >> /*@ lemma sq_double: \forall real x; x*x - .2 * x + 0.01 >= 0.; */
>     >
>     > This lemma is definitely wrong in WP with Real model (not float
>     there).
>     > At least, it is not provable in the forthcoming release of Frama-C.
>     > Is there a bug in some existing release?
>     >       L.
>     >
>     >
>     > _______________________________________________
>     > Frama-c-discuss mailing list
>     > Frama-c-discuss at lists.gforge.inria.fr
>     <mailto:Frama-c-discuss at lists.gforge.inria.fr>
>     > http://lists.gforge.inria.fr/mailman/listinfo/frama-c-discuss
>     >
>
>     --
>     Claude Marché                          | tel: +33 1 69 15 66 08
>     <tel:%2B33%201%2069%2015%2066%2008>
>     INRIA Saclay - ÃŽle-de-France           |
>     Université Paris-sud, Bat. 650         | http://www.lri.fr/~marche/
>     F-91405 <http://www.lri.fr/%7Emarche/%0AF-91405> ORSAY Cedex     
>                   |
>     _______________________________________________
>     Frama-c-discuss mailing list
>     Frama-c-discuss at lists.gforge.inria.fr
>     <mailto:Frama-c-discuss at lists.gforge.inria.fr>
>     http://lists.gforge.inria.fr/mailman/listinfo/frama-c-discuss
>
>
>
>
> -- 
> Boris
>
>
> _______________________________________________
> Frama-c-discuss mailing list
> Frama-c-discuss at lists.gforge.inria.fr
> http://lists.gforge.inria.fr/mailman/listinfo/frama-c-discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gforge.inria.fr/pipermail/frama-c-discuss/attachments/20160504/64a2f379/attachment.html>