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] Is it possible to "save"/reuse proofs?

mån 2019-04-15 klockan 12:15 +0200 skrev Virgile Prevosto:
> Hello Tomas,
> 
> > Le lun. 15 avr. 2019 à 11:36, Tomas Härdin <tjoppen at acc.umu.se> a écrit :
> 
> > 
> >   frama-c -wp -wp-rte -wp-timeout 30 -warn-unsigned-overflow -wp-prover
> > z3,cvc4,alt-ergo,qed proven.c
> > 
> > What am I missing?
> > 
> 
> As Loïc says, you're not missing anything obvious. However, there exists a
> hack that can allow you to take advantage of Why3's own session mechanism.
> Given the necessary steps, I'm afraid this is not exactly robust, but you
> might nevertheless be interested in the approach. This uses the GUI, but it
> might be possible to write a quick and dirty script that performs the file
> operations in batch mode.
> 
> 1) generate POs without launching provers: frama-c-gui [your options]
> -wp-prover none file.c
> 2) in the WP proofs panel of the GUI, launch Why3 on a PO: this should give
> you a Why3 session with all the PO attached to a given function.
> 3) Use whatever provers and strategies you want in Why3 and save the Why3
> session before quitting Why3 and Frama-C
> 4) copy the why3session.xml in ~/.frama-c-wp/project-session into somewhere
> safe
> 5) redo step 1
> 6) copy why3session.xml into ~/.frama-c-wp/project-session from the place
> where you stored it.
> 7) redo step 2: Why3 should read why3session.xml and reuse the proofs of
> the PO whose statement has not changed since step 3.

I've suspected something like this is possible, especially given -wp-
hints to feed coq scripts to WP. I haven't dived into either coq or
why3 much beyond looking at the output of -wp-out, but I suspect I
eventually will

/Tomas