[analyzer] exploded-graph-rewriter: NFC: Replace explorers with trimmers.

[NoQ]

Summary

Explorers aren't the right abstraction. For the purposes of displaying svg files we don't care in which order do we explore the nodes. We may care about this for other analyses, but we're not there yet.

The function of cutting out chunks of the graph is performed poorly by the explorers, because querying predecessors/successors on the explored nodes yields original successors/predecessors even if they aren't being explored.

Introduce a new entity, "trimmers", that do one thing but to it right: cut out chunks of the graph. Trimmers mutate the graph, so stale edges aren't even visible to their consumers in the pipeline. Additionally, trimmers are intrinsically composable: multiple trimmers can be applied to the graph sequentially.

Refactor the single-path explorer into the single-path trimmer. Rename the test file for consistency.

[Charusso]

In my mind an explorer would trim me a graph so I like the new abstraction to differentiate the two. Thanks!

clang/utils/analyzer/exploded-graph-rewriter.py

892 ↗

(On Diff #211984)

from the root. duplicated from the above line.

[NoQ]

Fxd.

cfe/trunk/test/Analysis/exploded-graph-rewriter/explorers.dot

-// RUN: %exploded_graph_rewriter %s \
-// RUN:     | FileCheck %s -check-prefixes=CHECK,BASIC
-// RUN: %exploded_graph_rewriter -s %s \
-// RUN:     | FileCheck %s -check-prefixes=CHECK,SINGLE
-
-// FIXME: Substitution doesn't seem to work on Windows.
-// UNSUPPORTED: system-windows
-
-Node0x1 [shape=record,label=
- "{{ "node_id": 1, "pointer": "0x1", "has_report": false, "is_sink": false,
-      "program_state": null, "program_points": []}\l}"];
-
-Node0x2 [shape=record,label=
- "{{ "node_id": 2, "pointer": "0x2", "has_report": false, "is_sink": false,
-      "program_state": null, "program_points": []}\l}"];
-
-Node0x3 [shape=record,label=
- "{{ "node_id": 3, "pointer": "0x3", "has_report": false, "is_sink": false,
-      "program_state": null, "program_points": []}\l}"];
-
-Node0x4 [shape=record,label=
- "{{ "node_id": 4, "pointer": "0x4", "has_report": false, "is_sink": false,
-      "program_state": null, "program_points": []}\l}"];
-
-// CHECK: Node0x1 -> Node0x2;
-Node0x1 -> Node0x2;
-
-// BASIC: Node0x1 -> Node0x3;
-// SINGLE-NOT: Node0x1 -> Node0x3;
-Node0x1 -> Node0x3;
-
-// CHECK: Node0x2 -> Node0x4;
-Node0x2 -> Node0x4;
-
-// BASIC: Node0x3 -> Node0x4;
-// SINGLE-NOT: Node0x3 -> Node0x4;
-Node0x3 -> Node0x4;

cfe/trunk/test/Analysis/exploded-graph-rewriter/trimmers.dot

+// RUN: %exploded_graph_rewriter %s \
+// RUN:     | FileCheck %s -check-prefixes=CHECK,BASIC
+// RUN: %exploded_graph_rewriter -s %s \
+// RUN:     | FileCheck %s -check-prefixes=CHECK,SINGLE
+
+// FIXME: Substitution doesn't seem to work on Windows.
+// UNSUPPORTED: system-windows
+
+Node0x1 [shape=record,label=
+ "{{ "node_id": 1, "pointer": "0x1", "has_report": false, "is_sink": false,
+      "program_state": null, "program_points": []}\l}"];
+
+Node0x2 [shape=record,label=
+ "{{ "node_id": 2, "pointer": "0x2", "has_report": false, "is_sink": false,
+      "program_state": null, "program_points": []}\l}"];
+
+Node0x3 [shape=record,label=
+ "{{ "node_id": 3, "pointer": "0x3", "has_report": false, "is_sink": false,
+      "program_state": null, "program_points": []}\l}"];
+
+Node0x4 [shape=record,label=
+ "{{ "node_id": 4, "pointer": "0x4", "has_report": false, "is_sink": false,
+      "program_state": null, "program_points": []}\l}"];
+
+// CHECK: Node0x1 -> Node0x2;
+Node0x1 -> Node0x2;
+
+// BASIC: Node0x1 -> Node0x3;
+// SINGLE-NOT: Node0x1 -> Node0x3;
+Node0x1 -> Node0x3;
+
+// CHECK: Node0x2 -> Node0x4;
+Node0x2 -> Node0x4;
+
+// BASIC: Node0x3 -> Node0x4;
+// SINGLE-NOT: Node0x3 -> Node0x4;
+Node0x3 -> Node0x4;

cfe/trunk/utils/analyzer/exploded-graph-rewriter.py

def explore(self, graph, visitor):
             logging.debug('Visiting ' + node)
             visitor.visit_node(graph.nodes[node])
             for succ in sorted(graph.nodes[node].successors):
                 logging.debug('Visiting edge: %s -> %s ' % (node, succ))
                 visitor.visit_edge(graph.nodes[node], graph.nodes[succ])
         visitor.visit_end_of_graph()
 
 
-# SinglePathExplorer traverses only a single path - the leftmost path
-# from the root. Useful when the trimmed graph is still too large
-# due to a large amount of equivalent reports.
-class SinglePathExplorer(object):
-    def __init__(self):
-        super(SinglePathExplorer, self).__init__()
+#===-----------------------------------------------------------------------===#
+# Trimmers cut out parts of the ExplodedGraph so that to focus on other parts.
+# Trimmers can be combined together by applying them sequentially.
+#===-----------------------------------------------------------------------===#
 
-    def explore(self, graph, visitor):
-        visitor.visit_begin_graph(graph)
 
-        # Keep track of visited nodes in order to avoid loops.
-        visited = set()
+# SinglePathTrimmer keeps only a single path - the leftmost path from the root.
+# Useful when the trimmed graph is still too large.
+class SinglePathTrimmer(object):
+    def __init__(self):
+        super(SinglePathTrimmer, self).__init__()
+
+    def trim(self, graph):
+        visited_nodes = set()
         node_id = graph.root_id
         while True:
-            visited.add(node_id)
+            visited_nodes.add(node_id)
             node = graph.nodes[node_id]
-            logging.debug('Visiting ' + node_id)
-            visitor.visit_node(node)
-            if len(node.successors) == 0:
-                break
-
+            if len(node.successors) > 0:
                 succ_id = node.successors[0]
                 succ = graph.nodes[succ_id]
-            logging.debug('Visiting edge: %s -> %s ' % (node_id, succ_id))
-            visitor.visit_edge(node, succ)
-            if succ_id in visited:
+                node.successors = [succ_id]
+                succ.predecessors = [node_id]
+                if succ_id in visited_nodes:
                     break
-
                 node_id = succ_id
-
-        visitor.visit_end_of_graph()
+            else:
+                break
+        graph.nodes = {node_id: graph.nodes[node_id]
+                       for node_id in visited_nodes}
 
 
 #===-----------------------------------------------------------------------===#
 # The entry point to the script.
 #===-----------------------------------------------------------------------===#
 
 
 def main():
def main():
     logging.basicConfig(level=args.loglevel)
 
     graph = ExplodedGraph()
     with open(args.filename) as fd:
         for raw_line in fd:
             raw_line = raw_line.strip()
             graph.add_raw_line(raw_line)
 
-    explorer = SinglePathExplorer() if args.single_path else BasicExplorer()
+    trimmers = []
+    if args.single_path:
+        trimmers.append(SinglePathTrimmer())
+
+    explorer = BasicExplorer()
+
     visitor = DotDumpVisitor(args.diff, args.dark, args.gray, args.topology,
                              args.dump_dot_only)
 
+    for trimmer in trimmers:
+        trimmer.trim(graph)
+
     explorer.explore(graph, visitor)
 
 
 if __name__ == '__main__':
     main()