[Soot-list] Using SOOT's dataflow classes for static numeric analysis that has infinite-height abstract domain and needs widening operator for termination

[Eric Bodden]

Hi Zell.

ForwardFlowAnalysis makes no assumptions about your domain. As long as
you guarantee termination, the analysis will just work. You just need
to implement flowThrough and merge in whatever way you like.

Eric


On 19 July 2012 05:44, Z <zell08v at orange.fr> wrote:
> Hello,
>
> Could you tell me please, how to use SOOT's dataflow framework to infer
> properties that can not be represented by the Flowset interface ?
>
> For example,  to implement a simple numeric analysis like interval analysis
> (which infers the possible interval range of each numeric variable at each
> program point), it seems no more possible to use  directly the generic
> ForwardAnalysis class, since the iteration strategy of the class assume that
> the height of the lattice is finite, whereas in most numeric analysis, the
> domain is not finite and the termination condition is ensured by a
> non-trivial widening operator.
>
> So if I understand correctly, one has to overload the doAnalysis of
> FowardAnalysis for writing a static numeric analyses?
>
> Thanks for your ideas.
> Zell.
>
>
> _______________________________________________
> Soot-list mailing list
> Soot-list at sable.mcgill.ca
> http://mailman.cs.mcgill.ca/mailman/listinfo/soot-list
>



-- 
Eric Bodden, Ph.D., http://bodden.de/
Head of Secure Software Engineering Group at EC SPRIDE
Principal Investigator in Secure Services at CASED
Tel: +49 6151 16-75422    Fax: +49 6151 16-72051
Room 3.2.14, Mornewegstr. 30, 64293 Darmstadt