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] Incremental verification

  • Subject: [Frama-c-discuss] Incremental verification
  • From: pascal.cuoq at (Pascal Cuoq)
  • Date: Mon, 4 Oct 2010 07:46:49 -0700
  • In-reply-to: <1286194386.7104.190.camel@iti27>
  • References: <1286194386.7104.190.camel@iti27>

Hello Boris,

> this is a proposal for a feature:  [...]
> I'd like to considerably reduce the time needed for Jessie in this loop.

This would be an interesting feature. Most of it can be implemented as
an independent Frama-C plug-in, without touching Jessie's internals.
The verification of a function must be run again if one of its callees' contract
has changed (not if the implementation of the callee has changed, though.
But then again, I think Jessie has a global analysis for aliases,
so you must not be using that feature for this reasoning to work).

All in all, you need at least a rudimentary parser for C+ACSL, but nothing too
sophisticated: being able to detect changes and to build the call tree
are enough.
It should also be possible to do it without using Frama-C and without using
OCaml as the implementation language.

Caching automated proof attempts from one run to the next would require
intervention inside the automated prover(s), though. Do some of them
already provide
this kind of feature (of course while preserving determinism of proof attempts:
success should never depend on the cache's contents)?