[Soot-list] Concrete field nodes exist in field-based points-to analysis in Spark

[Ondřej Lhoták]

Can you tell us the actual class name of the objects that you are
seeing? Can you provide some simple steps for us to reproduce what you
are seeing? Can you put a "throw new RuntimeException();" in the
constructor of the objects that you are seeing created, and send us
the stack trace that results when such an object is created?

On Thu, Apr 25, 2013 at 10:50:13AM -0400, Marc-André Laverdière-Papineau wrote:
> Hi,
> 
> Nobody replied yet. That looks like a question only Lhotak could answer ...
> 
> I am CCing him in case it helps :)
> 
> Marc-André Laverdière-Papineau
> Doctorant - PhD Candidate
> 
> On 04/22/2013 11:06 PM, Macro Angkor wrote:
> >Hi All,
> >
> > From my understanding, field-based points-to analysis in Spark groups
> >together field reference nodes that share the same field name and have
> >compatiable types for bases, and create a global VarNode for each group,
> >thus not touching concrete field nodes. However, with the phase option
> >"field-based" set to true, Spark's worklist propagator would still
> >introduce a lot of concrete field nodes. Anyone can help?
> >
> >Thanks in advance,
> >Michael
> >
> >
> >_______________________________________________
> >Soot-list mailing list
> >Soot-list at sable.mcgill.ca
> >http://mailman.cs.mcgill.ca/mailman/listinfo/soot-list
> >
>