Spark Pointer analysis

hi,

heres what i intend to do. I need a points to analysis graph at every program point and based on that i do a data flow analysis.

Now, when i generate the PAG as shown below, i get around 4757 alloc nodes which i understand are created from the standard class library.

Now, for a particular jimple statement i want to know

a. the nodes in the statement and

b. the nodes pointed to by these nodes.

for part a. i tried doing this :

If its an assignment statement,

Value lval = assignStmt.getLeftOp();

Value rval = assignStmt.getRightOp();

VarNode vLeftNode = findVarNode(lval);

VarNode vRightNode = findVarNode(rval);

System.out.println("VarNode (Right): ",+vRightNode.toString());
System.out.println("VarNode (Left): ",+vLeftNode.toString());

However, this gives me a null pointer exception. Am I doing something wrong here ??? and if that gets solved how do i solve part b.

Please help.

Regards,

Nikhil

PS: the code for construction of PAG

SootClass c = Scene.v().loadClassAndSupport("hello");
  c.setApplicationClass();
  Scene.v().setMainClass(c);

  Iterator itClasses=Scene.v().getClasses().iterator();
  while(itClasses.hasNext()){
     SootClass sc=(SootClass)itClasses.next();
     String scName=sc.getName();
   sc.setApplicationClass();
  }
  System.out.print("*** Constructing the PAG...");
  HashMap map = new HashMap();
  System.out.println("*** Reading default options ");
  String defaultOptions = SparkOptions.getDefaultOptions();
  StringTokenizer decoupe = new StringTokenizer(defaultOptions, " ");
  while(decoupe.hasMoreElements()){
   String option = decoupe.nextToken();
   StringTokenizer decoupeOption = new StringTokenizer(option, ":");
   String key = decoupeOption.nextToken();
   String value = decoupeOption.nextToken();
   map.put(key, value);
   System.out.println("Option "+key+"   * Set "+value);
  }
  SparkOptions options = new SparkOptions(map);

SparkTransformer.v().internalTransform("wjtp", map);
PAG pag = (PAG)Scene.v().getActivePointsToAnalysis();