[Soot-list] DeadAssignmentEliminator and assignments x=null
Chris Pickett
chris.pickett at mail.mcgill.ca
Tue Sep 11 17:15:25 EDT 2007
In fact, even System.gc() is just a suggestion.
Chris
Andrey_Poliektov at amat.com wrote:
>
> Hi Eric,
>
> As far I remember java spec, the result of your "test" is undetermined.
>
> for w.get() to return null, GC MUST determine that the object is weakly
> referenced, and this is not guaranteed to happen at any specific time.
>
> It does not matter how your optimization affects the test the result of
> which was undetermined to begin with.
>
> Thank you,
> Andrey
>
>
>
>
>
>
> *"Eric Bodden" <eric.bodden at mail.mcgill.ca>*
> Sent by: soot-list-bounces at sable.mcgill.ca
>
> 09/11/2007 01:54 PM
>
>
> To
> "Patrick Lam" <plam at sable.mcgill.ca>
> cc
> Soot list <soot-list at sable.mcgill.ca>
> Subject
> Re: [Soot-list] DeadAssignmentEliminator and assignments x=null
>
>
>
>
>
>
>
>
>
>
> > I don't think it's wrong to mark x=null as essential, but I don't think
> > it's bad to leave it as it is either. I'll leave it up to you.
>
> Point is that the current behavior forbids test cases like this:
>
> Object o = new Object();
> WeakReference w = new WeakReference(o);
> o = null; //eliminated
> System.gc();
> assert w.get()==null; //will fail
>
> Eric
>
> --
> Eric Bodden
> Sable Research Group
> McGill University, Montréal, Canada
> _______________________________________________
> Soot-list mailing list
> Soot-list at sable.mcgill.ca
> http://mailman.cs.mcgill.ca/mailman/listinfo/soot-list
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Soot-list mailing list
> Soot-list at sable.mcgill.ca
> http://mailman.cs.mcgill.ca/mailman/listinfo/soot-list
More information about the Soot-list
mailing list