Create in-memory cache to prevent duplicate reports

Author: bcyphers

src/js/background.js

@@ -1004,7 +1004,7 @@ Badger.prototype = {
   },
 
   /**
-   * Checks if local storage ( in dict) has any high-entropy keys
+   * Checks if local storage (in dict) has any high-entropy keys
    *
    * @param {Object} lsItems Local storage dict
    * @returns {boolean} true if it seems there are supercookies

src/js/constants.js

@@ -42,12 +42,15 @@ var exports = {
 
   // The max amount of time (in milliseconds) that PB will wait before sharing a
   // tracking action with EFF for community learning
-  MAX_CL_WAIT_TIME: 0, //30 * 60 * 1000, // one half hour
+  MAX_CL_WAIT_TIME: 5 * 60 * 1000, // five minutes
 
   // The probability that any given tracking action will be logged to the
   // community server, as a float from 0.0 to 1.0
   CL_PROBABILITY: 1.0,
 
+  // size of the in-memory community learning cache
+  CL_CACHE_SIZE: 5000,
+
   DNT_POLICY_CHECK_INTERVAL: 1000, // one second
 };
 

src/js/heuristicblocking.js

@@ -37,6 +37,9 @@ function HeuristicBlocker(pbStorage) {
   // impossible to attribute to a tab.
   this.tabOrigins = {};
   this.tabUrls = {};
+
+  // in-memory cache for community learning
+  this.previouslySharedTrackers = new Set();
 }
 
 HeuristicBlocker.prototype = {
@@ -356,6 +359,29 @@ HeuristicBlocker.prototype = {
   shareTrackerInfo: function(page_host, tracker_host, tracker_type) {
     // Share a random sample of trackers we observe
     if (Math.random() < constants.CL_PROBABILITY) {
+      // check if we've shared this tracker recently
+      // note that this check comes after checking against the snitch map
+      let tr_str = page_host + '+' + tracker_host + '+' + tracker_type;
+      if (this.previouslySharedTrackers.has(tr_str)) {
+        return;
+      }
+
+      // add this entry to the cache
+      this.previouslySharedTrackers.add(tr_str);
+
+      // if the cache gets too big, cut it in half
+      if (this.previouslySharedTrackers.size > constants.CL_CACHE_SIZE) {
+        this.previouslySharedTrackers = new Set(
+          // An array created from the set will have all of its entries ordered
+          // by when they were added
+          Array.from(this.previouslySharedTrackers).slice(
+            // keep the most recent half of the cache entries
+            Math.floor(constants.CL_CACHE_SIZE / 2)
+          )
+        );
+      }
+
+      // now make the request to the database server
       setTimeout(function() {
        fetch("http://localhost:8080", {
          method: "POST",