mononoke: walker: filter the repo path by node type

Summary:
Not all node types can have a path associated

Reset the tracked path to None if the route is taking us through a node type that can't have a repo path.

Reviewed By: krallin

Differential Revision: D21228372

fbshipit-source-id: 2b1e291f09232500adce79c630d428f09cd2d2cc

eden/mononoke/tests/integration/test-walker-sample-by-path.t

@@ -62,23 +62,23 @@ Three separate cycles allowing all edges, total bytes should be the same as full
   * Run */s,*/s,579,579,0%,*s; * (glob)
   Walked/s,* (glob)
 
-Reduced edge three separate cycles moving offset each time, total in each cycle should be the same as above. Its not.
+Reduced edge three separate cycles moving offset each time, total in each cycle should be the same as above.
   $ for i in {0..2}; do mononoke_walker --storage-id=blobstore --readonly-storage compression-benefit -q --bookmark master_bookmark -I deep -x BonsaiFsnodeMapping -x Fsnode -X BonsaiChangesetToBonsaiParent -X HgFileEnvelopeToFileContent -X HgChangesetToHgParent --sample-rate=3 --sample-offset=$i 2>&1; done | strip_glog
   Walking roots * (glob)
   Walking edge types * (glob)
   Walking node types * (glob)
   Final count: * (glob)
-  * Run */s,*/s,1045,1016,2%,*s; * (glob)
+  * Run */s,*/s,1146,1117,2%,*s; * (glob)
   Walked/s,* (glob)
   Walking roots * (glob)
   Walking edge types * (glob)
   Walking node types * (glob)
   Final count: * (glob)
-  * Run */s,*/s,522,522,0%,*s;* (glob)
+  * Run */s,*/s,443,443,0%,*s;* (glob)
   Walked/s,* (glob)
   Walking roots * (glob)
   Walking edge types * (glob)
   Walking node types * (glob)
   Final count: * (glob)
-  * Run */s,*/s,601,601,0%,*s; * (glob)
+  * Run */s,*/s,579,579,0%,*s; * (glob)
   Walked/s,* (glob)

eden/mononoke/walker/src/sampling.rs

@@ -47,6 +47,32 @@ pub struct PathTrackingRoute {
     pub path: Option<WrappedPath>,
 }
 
+// Only certain node types can have repo paths associated
+fn filter_repo_path(node_type: NodeType, path: Option<&'_ WrappedPath>) -> Option<&'_ WrappedPath> {
+    match node_type {
+        NodeType::Root => None,
+        // Bonsai
+        NodeType::Bookmark => None,
+        NodeType::BonsaiChangeset => None,
+        NodeType::BonsaiHgMapping => None,
+        NodeType::BonsaiPhaseMapping => None,
+        NodeType::PublishedBookmarks => None,
+        NodeType::BonsaiFsnodeMapping => None,
+        // Hg
+        NodeType::HgBonsaiMapping => None,
+        NodeType::HgChangeset => None,
+        NodeType::HgManifest => path,
+        NodeType::HgFileEnvelope => path,
+        NodeType::HgFileNode => path,
+        // Content
+        NodeType::FileContent => path,
+        NodeType::FileContentMetadata => path,
+        NodeType::AliasContentMapping => path,
+        // Derived Data
+        NodeType::Fsnode => path,
+    }
+}
+
 impl PathTrackingRoute {
     fn evolve_path<'a>(
         from_route: Option<&'a WrappedPath>,
@@ -59,8 +85,8 @@ impl PathTrackingRoute {
             None => match target.stats_path() {
                 // Path is part of node identity
                 Some(from_node) => Some(from_node),
-                // No per-node path, so use the path from route
-                None => from_route,
+                // No per-node path, so use the route, filtering out nodes that can't have repo paths
+                None => filter_repo_path(target.get_type(), from_route),
             },
         }
     }