chimera: hybrid of Hyperscan and PCRE

tools/hscollider/ResultSet.h

@@ -35,25 +35,36 @@
 #include <utility>
 #include <vector>
 
+// Type for capturing groups: a vector of (from, to) offsets, with both set to
+// -1 for inactive groups (like pcre's ovector). Used by hybrid modes.
+typedef std::vector<std::pair<int, int> > CaptureVec;
+
 // Class representing a single match, encapsulating to/from offsets.
 class MatchResult {
 public:
     MatchResult(unsigned long long start, unsigned long long end)
         : from(start), to(end) {}
+    MatchResult(unsigned long long start, unsigned long long end,
+                const CaptureVec &cap)
+            : from(start), to(end), captured(cap) {}
 
     bool operator<(const MatchResult &a) const {
         if (from != a.from) {
             return from < a.from;
         }
-        return to < a.to;
+        if (to != a.to) {
+            return to < a.to;
+        }
+        return captured < a.captured;
     }
 
     bool operator==(const MatchResult &a) const {
-        return from == a.from && to == a.to;
+        return from == a.from && to == a.to && captured == a.captured;
     }
 
     unsigned long long from;
     unsigned long long to;
+    CaptureVec captured;
 };
 
 enum ResultSource {
@@ -114,6 +125,19 @@ public:
         }
     }
 
+    // Add a match (with capturing vector)
+    void addMatch(unsigned long long from, unsigned long long to,
+                  const CaptureVec &cap, int block = 0) {
+        MatchResult m(from, to, cap);
+        matches.insert(m);
+
+        if (matches_by_block[block].find(m) != matches_by_block[block].end()) {
+            dupe_matches.insert(m);
+        } else {
+            matches_by_block[block].insert(m);
+        }
+    }
+
     // Clear all matches.
     void clear() {
         matches.clear();