Merge #76

76: hyperloglog: fix potential array out-of-bounds in bias correction r=crepererum a=crepererum

Fixes #74

Co-authored-by: Marco Neumann <marco@crepererum.net>

src/hyperloglog.rs

@@ -175,8 +175,15 @@ where
         // binary search first nearest neighbor
         let lookup_array = RAW_ESTIMATE_DATA_VEC[self.b - RAW_ESTIMATE_DATA_OFFSET];
         let mut idx_left = match lookup_array.binary_search_by(|v| v.partial_cmp(&e).unwrap()) {
-            Ok(i) => Some(i),  // exact match
-            Err(i) => Some(i), // no match, i points to left neighbor
+            Ok(i) => Some(i), // exact match
+            Err(i) => {
+                // no match, i points to left neighbor
+                if i < lookup_array.len() {
+                    Some(i) // neighbor is in-range
+                } else {
+                    Some(i - 1) // neighbor at the end of the array
+                }
+            }
         };
 
         let mut idx_right = match idx_left {
@@ -586,4 +593,73 @@ mod tests {
         assert_eq!(hll.count(), 1);
         assert!(!hll.is_empty());
     }
+
+    #[test]
+    fn issue_74() {
+        let panic_data = vec![
+            "ofr-1-1517560282779878449",
+            "ofr-1-1517589543534331019",
+            "ofr-1-1517590532450550786",
+            "ofr-1-1517644560121333465",
+            "ofr-1-1517746611185649116",
+            "ofr-1-1518051376300950677",
+            "ofr-1-1518484387459892414",
+            "ofr-1-1518488008830355319",
+            "ofr-1-1518488407814571264",
+            "ofr-1-1518561818180978525",
+            "ofr-1-1518678274740717330",
+            "ofr-1-1519461045930165638",
+            "ofr-1-1519470647696557288",
+            "ofr-1-1519567114956309703",
+            "ofr-1-1519653616441755584",
+            "ofr-1-1519655049912256356",
+            "ofr-1-1520105514088138521",
+            "ofr-1-1520294225822221822",
+            "ofr-1-1520319017418884884",
+            "ofr-1-1520505982893295286",
+            "ofr-1-1520553027150677707",
+            "ofr-1-1520925550686111649",
+            "ofr-1-1520927095122167663",
+            "ofr-1-1521290010424640726",
+            "ofr-1-1521458659554886917",
+            "ofr-1-1521943577454052994",
+            "ofr-1-1521971260753839540",
+            "ofr-1-1522000670785668758",
+            "ofr-1-1522043914876749176",
+            "ofr-1-1522206531944580201",
+            "ofr-1-1522234960069920034",
+            "ofr-1-1522333169901504119",
+            "ofr-1-1522363887846294936",
+            "ofr-1-1522484446749918495",
+            "ofr-1-1522600458059122179",
+            "ofr-1-1522687450205783676",
+            "ofr-1-1522765602785461678",
+            "ofr-1-1522815395559769187",
+            "ofr-1-1522839112893465736",
+            "ofr-1-1523001178903151627",
+            "ofr-1-1523018056414397988",
+            "ofr-1-1523096555609261412",
+            "ofr-1-1523103371222189143",
+            "ofr-1-1523256333918667890",
+            "ofr-1-1523270427746895732",
+            "ofr-1-1523411745695466681",
+            "ofr-1-1523630566301631536",
+            "ofr-1-1523839014553388093",
+            "ofr-1-1523894230803940925",
+            "ofr-1-1523931915564221543",
+            "ofr-1-1524104734332815100",
+            "ofr-1-1524113364834715372",
+            "ofr-1-1524209603273164167",
+            "ofr-1-1524276802153219312",
+            "ofr-1-1524554894791804305",
+            "ofr-1-1524621894100584193",
+        ];
+
+        let mut hll = HyperLogLog::new(4);
+        for entry in &panic_data {
+            hll.add(entry);
+        }
+
+        hll.count();
+    }
 }