[Lookalikes] Ignore edit distance matches that only differ with a "-"

This CL marks edit distance matches that only differ due to a single
dash. This is a temporary mitigation for a large class of false
positives until a DAL approach is able to offer better tools for domain
owners.

Fixed: 1159473
Change-Id: Ibe2f34b71ee219a01383b0d46937d2749ba50a5b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2628073
Commit-Queue: Joe DeBlasio <jdeblasio@chromium.org>
Reviewed-by: Mustafa Emre Acer <meacer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#843257}

components/lookalikes/core/lookalike_url_util.cc

@@ -347,6 +347,22 @@ bool IsAllowedToBeEmbedded(
          EndsWithPermittedDomains(embedded_target, embedding_domain);
 }
 
+// Returns the first character of the first string that is different from the
+// second string. Strings should be at least 1 edit distance apart.
+char GetFirstDifferentChar(const std::string& str1, const std::string& str2) {
+  std::string::const_iterator i1 = str1.begin();
+  std::string::const_iterator i2 = str2.begin();
+  while (i1 != str1.end() && i2 != str2.end()) {
+    if (*i1 != *i2) {
+      return *i1;
+    }
+    i1++;
+    i2++;
+  }
+  NOTREACHED();
+  return 0;
+}
+
 }  // namespace
 
 DomainInfo::DomainInfo(const std::string& arg_hostname,
@@ -497,6 +513,17 @@ bool IsLikelyEditDistanceFalsePositive(const DomainInfo& navigated_domain,
     }
   }
 
+  // Ignore domains that only differ by an insertion of a "-".
+  if (nav_dom_len != matched_dom_len) {
+    if (nav_dom_len < matched_dom_len &&
+        GetFirstDifferentChar(matched_dom, nav_dom) == '-') {
+      return true;
+    } else if (nav_dom_len > matched_dom_len &&
+               GetFirstDifferentChar(nav_dom, matched_dom) == '-') {
+      return true;
+    }
+  }
+
   return false;
 }
 

components/lookalikes/core/lookalike_url_util.h

@@ -138,7 +138,7 @@ bool IsEditDistanceAtMostOne(const base::string16& str1,
 // Returns whether |navigated_domain| and |matched_domain| are likely to be edit
 // distance false positives, and thus the user should *not* be warned.
 //
-// Assumes |navigated_domain| and |matched_domain| are edit distance matches.
+// Assumes |navigated_domain| and |matched_domain| are edit distance of 1 apart.
 bool IsLikelyEditDistanceFalsePositive(const DomainInfo& navigated_domain,
                                        const DomainInfo& matched_domain);
 

components/lookalikes/core/lookalike_url_util_unittest.cc

@@ -105,6 +105,14 @@ TEST(LookalikeUrlUtilTest, EditDistanceExcludesCommonFalsePositives) {
       {"abcde.com", "axbcde.com", false},   // Deletion
       {"axbcde.com", "abcde.com", false},   // Insertion
       {"axbcde.com", "aybcde.com", false},  // Substitution
+
+      // We permit matches that only differ due to a single "-".
+      {"-abcde.com", "abcde.com", true},
+      {"ab-cde.com", "abcde.com", true},
+      {"abcde-.com", "abcde.com", true},
+      {"abcde.com", "-abcde.com", true},
+      {"abcde.com", "ab-cde.com", true},
+      {"abcde.com", "abcde-.com", true},
   };
   for (const TestCase& test_case : kTestCases) {
     auto navigated =