Python: better (if imperfect) handling of phi node

Author: yoff

python/ql/src/experimental/dataflow/internal/DataFlowPrivate.qll

@@ -75,7 +75,7 @@ module EssaFlow {
     or
     exists(PhiFunction p |
       nodeTo.asEssaNode() = p.getVariable() and
-      nodeFrom.asEssaNode() = p.getShortCircuitInput()
+      nodeFrom.asEssaNode() = p.getAnInput()
     )
   }
 }

python/ql/test/experimental/dataflow/regression/dataflow.expected

@@ -4,6 +4,8 @@
 | test.py:10:12:10:17 | ControlFlowNode for SOURCE | test.py:17:10:17:10 | ControlFlowNode for t |
 | test.py:20:9:20:14 | ControlFlowNode for SOURCE | test.py:13:10:13:12 | ControlFlowNode for arg |
 | test.py:37:13:37:18 | ControlFlowNode for SOURCE | test.py:41:14:41:14 | ControlFlowNode for t |
+| test.py:62:13:62:18 | ControlFlowNode for SOURCE | test.py:13:10:13:12 | ControlFlowNode for arg |
+| test.py:67:13:67:18 | ControlFlowNode for SOURCE | test.py:13:10:13:12 | ControlFlowNode for arg |
 | test.py:76:9:76:14 | ControlFlowNode for SOURCE | test.py:78:10:78:10 | ControlFlowNode for t |
 | test.py:108:13:108:18 | ControlFlowNode for SOURCE | test.py:112:14:112:14 | ControlFlowNode for t |
 | test.py:139:10:139:15 | ControlFlowNode for SOURCE | test.py:140:14:140:14 | ControlFlowNode for t |