C++: Faster implementation of BB entry node

jbj · jbj · commit 265852058d61 · 2018-10-05T14:20:03.000+02:00
The existing implementation of `primitive_basic_block_entry_node` was
"cleverly" computing two properties about `node` with a single
`strictcount`: whether `node` had multiple predecessors and whether any
of those predecessors had more than once successor. This was fast enough
on most snapshots, but on the snapshot of our own code it took 37
seconds to compute `primitive_basic_block_entry_node` and its auxiliary
predicates. This is likely to have affected other large snapshots too.

With this change, the property is computed like in our other languages,
and it brings the run time down to 4 seconds.
diff --git a/cpp/ql/src/semmle/code/cpp/controlflow/internal/PrimitiveBasicBlocks.qll b/cpp/ql/src/semmle/code/cpp/controlflow/internal/PrimitiveBasicBlocks.qll
@@ -31,8 +31,11 @@ private cached module Cached {
     // or the node's predecessor has more than one successor,
     // then the node is the start of a new primitive basic block.
     or
-    strictcount (Node pred, Node other
-    | successors_extended(pred,node) and successors_extended(pred,other)) > 1
+    strictcount(Node pred | successors_extended(pred, node)) > 1
+    or
+    exists(ControlFlowNode pred | successors_extended(pred, node) |
+      strictcount(ControlFlowNode other | successors_extended(pred, other)) > 1
+    )
 
     // If the node has zero predecessors then it is the start of
     // a BB. However, the C++ AST contains many nodes with zero
